| Title: | Harmonize Eye-Tracking, Pupillometry, Biometrics, and Psychometric Process Data |
|---|---|
| Description: | Provides an extensible, vendor-neutral framework for importing, validating, harmonizing, transforming, visualizing, and modelling eye-tracking, pupillometry, behavioural, and biometric process data. The package uses explicit timebase and coordinate-space registries, preserves native fields and provenance, and offers first-class adapters for Gazepoint Analysis and Gazepoint Biometrics exports alongside generic and vendor-specific importers. Downstream tools support trial and area of interest reconstruction, signal-quality auditing, feature derivation, scanpath analysis, response-time and item-response workflows, and optional psychometric modelling engines. An integrated Gazepoint workflow produces quality-control evidence, media-trial reconstruction, plots, analysis-ready process tables, item response theory (IRT)-ready response structures, and reproducible reports. Brain Imaging Data Structure (BIDS) interoperability for eye-tracking and validation-release infrastructure support disk-backed storage, independent multi-vendor evidence, grouped validation, simulation calibration, model-equivalence audits, and explicitly experimental advanced psychometric process models. Research-scale infrastructure adds deterministic resumable Monte Carlo execution, atomic validation checkpoints, explicit advanced-model promotion gates, independent multi-vendor evidence registries, stable object contracts, partitioned disk-backed storage, optional probabilistic engines, and a fully synthetic multimodal benchmark for reproducibility testing. The measurement-intelligence programme adds probabilistic and compositional area of interest (AOI) analysis, measurement-uncertainty propagation, calibration and device-transportability audits, process reliability, phase-amplitude pupil registration, informative-missingness sensitivity, right-censored gaze-latency survival analysis, temporal and spatial process models, item-bank decision optimization, fairness monitoring, conditional process reference distributions, and evidence-provenance graphs. |
| Authors: | Stefanos Balaskas [aut, cre] (ORCID: <https://orcid.org/0000-0003-2444-9796>) |
| Maintainer: | Stefanos Balaskas <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.12.0.9000 |
| Built: | 2026-10-09 21:30:29 UTC |
| Source: | https://github.com/stefanosbalaskas/eyeprocess |
An extensible, vendor-neutral framework for importing, validating, harmonizing, transforming, visualizing, and modelling eye-tracking, pupillometry, behavioural, and biometric process data, with first-class Gazepoint support.
The package represents heterogeneous exports as linked canonical tables with explicit timebase, coordinate-space, processing-level, quality, and provenance information. It supports generic mappings and dedicated adapters, then provides common downstream workflows for preprocessing, AOIs, trials, features, plots, and optional psychometric modelling.
Stefanos Balaskas
Package documentation and the methodological references cited in the vignettes.
Create formal channel-ablation datasets
ablate_multimodal_channels( x, include = c("response", "rt", "gaze", "pupil"), include_response = TRUE )ablate_multimodal_channels( x, include = c("response", "rt", "gaze", "pupil"), include_response = TRUE )
x |
An 'eye_multimodal_measurement'. |
include |
Character vector of channels eligible for ablation. |
include_response |
Whether response must remain in every scenario. |
An 'eye_multimodal_ablation' list.
Compute aDDM/GLAM-inspired gaze-evidence proxy features
addm_glam_proxy_features( data, by = c("person_id", "trial_id"), time = "time_ms", aoi = "aoi", target_aoi = "target", distractor_aoi = "distractor", action_aoi = "button" )addm_glam_proxy_features( data, by = c("person_id", "trial_id"), time = "time_ms", aoi = "aoi", target_aoi = "target", distractor_aoi = "distractor", action_aoi = "button" )
data |
Sample-level data. |
by |
Grouping columns. |
time, aoi
|
Columns. |
target_aoi, distractor_aoi, action_aoi
|
AOI labels. |
An object of class "eye_decision_process_proxy", stored as a named list, with components "features", "by", "status", "caveat". It contains aDDM/GLAM-inspired gaze-evidence proxy features and associated metadata or diagnostics needed to interpret the result.
Extract confound-adjusted pupil values
adjust_pupil_confounds(x)adjust_pupil_confounds(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing the fitted model data, including 'pupil_confound_residual' and 'pupil_confound_adjusted' columns.
Specify evidence required to promote advanced model interfaces
advanced_model_evidence_spec( models = c("fit_joint_process_model", "fit_shared_process_factor", "fit_strategy_mixture", "fit_process_irt", "fit_pupil_informed_irt", "fit_multimodal_irt", "fit_dynamic_aoi_model", "fit_gaze_weighted_choice", "fit_dynamic_irtree", "fit_joint_functional_pupil_irt", "fit_theory_strategy_irt", "fit_gaze_diffusion_irt"), require_recovery = TRUE, require_calibration = TRUE, require_misspecification = TRUE, require_grouped_validation = TRUE, require_engine_equivalence = TRUE, require_empirical_reproduction = TRUE, require_sensitivity = TRUE )advanced_model_evidence_spec( models = c("fit_joint_process_model", "fit_shared_process_factor", "fit_strategy_mixture", "fit_process_irt", "fit_pupil_informed_irt", "fit_multimodal_irt", "fit_dynamic_aoi_model", "fit_gaze_weighted_choice", "fit_dynamic_irtree", "fit_joint_functional_pupil_irt", "fit_theory_strategy_irt", "fit_gaze_diffusion_irt"), require_recovery = TRUE, require_calibration = TRUE, require_misspecification = TRUE, require_grouped_validation = TRUE, require_engine_equivalence = TRUE, require_empirical_reproduction = TRUE, require_sensitivity = TRUE )
models |
Advanced model function names. |
require_recovery |
Require passing parameter-recovery evidence. |
require_calibration |
Require simulation-based calibration evidence. |
require_misspecification |
Require evidence that prespecified failure scenarios are detected. |
require_grouped_validation |
Require grouped out-of-sample validation. |
require_engine_equivalence |
Require comparison with a benchmark engine. |
require_empirical_reproduction |
Require a licensed empirical reproduction. |
require_sensitivity |
Require a multi-specification preprocessing/AOI sensitivity analysis. |
An 'eye_advanced_evidence_spec'.
By default this returns a one-factor-at-a-time screening design around a declared reference scenario. This preserves every factor and level from the research programme without accidentally launching hundreds of thousands of Monte Carlo scenarios. Set 'full_factorial = TRUE' only when the computing plan explicitly supports the complete Cartesian design.
advanced_validation_grid(quick = FALSE, full_factorial = FALSE)advanced_validation_grid(quick = FALSE, full_factorial = FALSE)
quick |
Whether to return a compact smoke-test design. |
full_factorial |
Whether to return the complete Cartesian design. |
A scenario data frame for 'run_model_validation()' or custom Monte Carlo programmes.
Extract algorithm facet effects
algorithm_facet_effects(object, channel = c("response", "process"))algorithm_facet_effects(object, channel = c("response", "process"))
object |
A fitted eyeprocess model or audit object. |
channel |
Measurement channel to inspect. |
An object of class "eye_process_facet_effects", stored as a named list, with components "facet", "column", "channel", "random_effects", "variance_component". It contains algorithm facet effects and associated metadata or diagnostics needed to interpret the result.
Entropy of an explicitly enumerated analysis-decision space
analysis_decision_entropy(..., base = 2)analysis_decision_entropy(..., base = 2)
... |
Named option vectors. |
base |
Logarithm base. |
Data frame with option counts and maximum entropy under equal weighting.
Snapshot an eyeprocess analysis environment
analysis_environment_snapshot(packages = loadedNamespaces())analysis_environment_snapshot(packages = loadedNamespaces())
packages |
Optional package names; defaults to loaded namespaces. |
A named list with components "r_version", "platform", "os", "locale", "timezone", "packages", containing snapshot an eyeprocess analysis environment and associated metadata or diagnostics.
Audit compatibility between measurement resolution and an analysis target
analysis_resolution_guard( event_duration_ms, effective_hz, spatial_feature_size = NA_real_, radial_error = NA_real_, min_samples = 3, max_error_fraction = 0.5 )analysis_resolution_guard( event_duration_ms, effective_hz, spatial_feature_size = NA_real_, radial_error = NA_real_, min_samples = 3, max_error_fraction = 0.5 )
event_duration_ms |
Smallest event duration the analysis intends to resolve. |
effective_hz |
Empirical sampling frequency. |
spatial_feature_size |
Optional smallest spatial feature/AOI dimension in coordinate units. |
radial_error |
Optional empirical radial error in the same spatial units. |
min_samples |
User-declared minimum samples per temporal feature. |
max_error_fraction |
User-declared maximum spatial-error / feature-size ratio. |
A named list with components "expected_samples", "temporal_ok", "spatial_error_fraction", "spatial_ok", "min_samples", "max_error_fraction", "overall", "caveat", containing compatibility between measurement resolution and an analysis target and associated metadata or diagnostics.
AOI membership probabilities from uncertainty draws
aoi_membership_probability(draws, aois)aoi_membership_probability(draws, aois)
draws |
Output of 'propagate_calibration_uncertainty()' or compatible table. |
aois |
Rectangular AOI table with aoi/x_min/x_max/y_min/y_max. |
A tabular R object containing aOI membership probabilities from uncertainty draws; rows represent analysis units and columns contain the returned quantities.
Vendor-neutral tools for treating area-of-interest (AOI) geometry as an explicit analytical assumption. The workflow validates rectangle and polygon geometry, constructs reproducible perturbations, remaps gaze or fixation coordinates, recomputes AOI features, optionally reruns a user-supplied statistical model, and reports assignment and inference stability without interpreting robustness frequencies as probabilities that a scientific conclusion is true.
validate_aoi_geometry(aois, allow_overlap = TRUE) convert_aoi_margin_to_degrees( margin_pixels, screen_width_px, screen_height_px, viewing_distance, physical_screen_size ) convert_aoi_margin_to_pixels( margin_degrees, screen_width_px, screen_height_px, viewing_distance, physical_screen_size ) aoi_perturbation_spec( perturbation_id, operation = "baseline", margin_x = 0, margin_y = margin_x, translation_x = 0, translation_y = 0, unit = "px", screen_width_px = NULL, screen_height_px = NULL, viewing_distance = NULL, physical_screen_size = NULL, degrees_per_pixel = NULL, seed = NULL, boundary_policy = "warn" ) dilate_aoi(aois, margin, ...) erode_aoi(aois, margin, ...) translate_aoi(aois, x = 0, y = 0, ...) jitter_aoi(aois, x, y = x, seed = 20260918L, ...) perturb_aoi_geometry(aois, spec) create_aoi_perturbation_grid( dilations = NULL, erosions = NULL, translations_x = NULL, translations_y = NULL, translations_xy = NULL, jitters = NULL, anisotropic = NULL, unit = "px", include_baseline = TRUE, seed = 20260918L, ... ) apply_aoi_perturbation_grid(aois, grid) compare_aoi_assignments(baseline, perturbed, ids = seq_along(baseline)) estimate_aoi_assignment_stability( comparisons, metadata = NULL, group_cols = NULL ) estimate_fixation_assignment_probability( assignments, ids = NULL, include_baseline = TRUE ) recompute_aoi_features( data, assignments, participant_col = NULL, trial_col = NULL, duration_col = NULL, time_col = NULL, perturbation_id = NULL, aoi_levels = NULL, observation_level = c("fixation", "sample") ) run_aoi_sensitivity_analysis( data, aois, grid, x_col, y_col, observation_id_col = NULL, participant_col = NULL, trial_col = NULL, duration_col = NULL, time_col = NULL, observation_level = c("fixation", "sample"), overlap_policy = "ambiguous", model_callback = NULL, preprocessing_specification = NULL, event_detector = NULL, quality_rules = NULL, model_specification = NULL ) assess_aoi_inference_stability(x, term = NULL) summarise_aoi_sensitivity(x) report_aoi_sensitivity(x) plot_aoi_perturbations( x, perturbation_id = NULL, data = NULL, x_col = NULL, y_col = NULL, ... ) plot_aoi_assignment_stability(x, ...) plot_aoi_coefficient_stability(x, term, ...) plot_aoi_robustness_surface( x, value_col = "proportion_unchanged", x_col = "margin_x", y_col = "margin_y", ... )validate_aoi_geometry(aois, allow_overlap = TRUE) convert_aoi_margin_to_degrees( margin_pixels, screen_width_px, screen_height_px, viewing_distance, physical_screen_size ) convert_aoi_margin_to_pixels( margin_degrees, screen_width_px, screen_height_px, viewing_distance, physical_screen_size ) aoi_perturbation_spec( perturbation_id, operation = "baseline", margin_x = 0, margin_y = margin_x, translation_x = 0, translation_y = 0, unit = "px", screen_width_px = NULL, screen_height_px = NULL, viewing_distance = NULL, physical_screen_size = NULL, degrees_per_pixel = NULL, seed = NULL, boundary_policy = "warn" ) dilate_aoi(aois, margin, ...) erode_aoi(aois, margin, ...) translate_aoi(aois, x = 0, y = 0, ...) jitter_aoi(aois, x, y = x, seed = 20260918L, ...) perturb_aoi_geometry(aois, spec) create_aoi_perturbation_grid( dilations = NULL, erosions = NULL, translations_x = NULL, translations_y = NULL, translations_xy = NULL, jitters = NULL, anisotropic = NULL, unit = "px", include_baseline = TRUE, seed = 20260918L, ... ) apply_aoi_perturbation_grid(aois, grid) compare_aoi_assignments(baseline, perturbed, ids = seq_along(baseline)) estimate_aoi_assignment_stability( comparisons, metadata = NULL, group_cols = NULL ) estimate_fixation_assignment_probability( assignments, ids = NULL, include_baseline = TRUE ) recompute_aoi_features( data, assignments, participant_col = NULL, trial_col = NULL, duration_col = NULL, time_col = NULL, perturbation_id = NULL, aoi_levels = NULL, observation_level = c("fixation", "sample") ) run_aoi_sensitivity_analysis( data, aois, grid, x_col, y_col, observation_id_col = NULL, participant_col = NULL, trial_col = NULL, duration_col = NULL, time_col = NULL, observation_level = c("fixation", "sample"), overlap_policy = "ambiguous", model_callback = NULL, preprocessing_specification = NULL, event_detector = NULL, quality_rules = NULL, model_specification = NULL ) assess_aoi_inference_stability(x, term = NULL) summarise_aoi_sensitivity(x) report_aoi_sensitivity(x) plot_aoi_perturbations( x, perturbation_id = NULL, data = NULL, x_col = NULL, y_col = NULL, ... ) plot_aoi_assignment_stability(x, ...) plot_aoi_coefficient_stability(x, term, ...) plot_aoi_robustness_surface( x, value_col = "proportion_unchanged", x_col = "margin_x", y_col = "margin_y", ... )
aois |
A data frame containing rectangular or polygonal AOI geometry. |
allow_overlap |
Whether overlapping AOIs are allowed during validation. |
margin_pixels, margin_degrees, margin, margin_x, margin_y
|
Perturbation margins in the declared unit. |
screen_width_px, screen_height_px
|
Declared display resolution in pixels. |
viewing_distance |
Viewing distance in the same physical unit used for screen size. |
physical_screen_size |
Length-two physical screen width and height. |
perturbation_id |
Stable identifier for a perturbation branch. |
operation |
One of baseline, dilation, erosion, translate, jitter, or anisotropic expansion. |
translation_x, translation_y, y
|
Horizontal and vertical translation values. |
unit |
Either pixels (px) or degrees of visual angle (deg). |
degrees_per_pixel |
Optional explicit horizontal and vertical degrees-per-pixel values. |
seed |
Reproducible random seed used for jitter. |
boundary_policy |
One of warn, clip, error, or allow for geometry outside the declared screen. |
spec |
An AOI perturbation specification. |
dilations, erosions, translations_x, translations_y, translations_xy, jitters, anisotropic
|
Vectors or pairs defining a deterministic perturbation grid. 'translations_xy' accepts a length-two vector, two-column matrix/data frame, or list of x/y pairs. |
include_baseline |
Whether to include the unchanged nominal branch. |
grid |
An AOI perturbation grid. |
baseline, perturbed
|
Nominal and perturbed AOI assignment vectors. |
ids |
Optional observation identifiers. |
comparisons |
Named AOI assignment comparison objects. |
metadata |
Optional one-row-per-observation metadata. |
group_cols |
Optional participant, trial, or other grouping columns. |
assignments |
AOI assignment vector or named branch list, depending on the function. |
data |
Sample- or fixation-level data. |
participant_col, trial_col
|
Optional participant and trial grouping columns. |
duration_col, time_col
|
Optional duration and time columns used for AOI feature recomputation. |
aoi_levels |
Optional complete AOI identifier universe. When supplied, observed participant/trial groups retain explicit zero-count AOI cells; groups with no valid assignment opportunity retain missing values rather than zeros. |
observation_level |
Explicit input level, either fixation or sample. Universal observation counts/timing are retained without labeling sample rows as fixations. |
x_col, y_col
|
Coordinate columns in data or robustness-surface dimensions. |
observation_id_col |
Optional unique observation identifier column. |
overlap_policy |
How observed overlapping membership is represented; defaults to explicit ambiguity. |
model_callback |
Optional function receiving recomputed features, assigned row data, and the current perturbation specification. |
preprocessing_specification, event_detector, quality_rules, model_specification
|
Provenance fields retained in the result object. |
x |
Horizontal translation or jitter value for geometry helpers, or an AOI sensitivity, perturbation, or stability object for plotting and summary methods, depending on the function. |
term |
Model term to summarize or plot. |
value_col |
Robustness quantity used in the surface plot. |
... |
Additional perturbation-specification or plotting arguments. |
Overlapping AOIs are not resolved silently. The default row-level assignment
for overlap is the explicit label __ambiguous__. Missing coordinates
remain missing. Feature recomputation distinguishes structural zero AOI cells from groups with no valid assignment opportunity; the latter remain missing. Degree-based perturbations require declared display geometry
and viewing distance, or explicit degrees-per-pixel values.
A model callback must explicitly choose and fit the desired estimator and return at least term, estimate, SE, CI_low, CI_high, p_value, model_converged, and N. Convergence must be explicit; converged rows require finite estimates, standard errors, confidence limits, and positive integer N. Callback errors and non-converged fits remain visible in the audit trail, and inference summaries retain the model-N range across converged perturbations.
Same-sign proportions and assignment frequencies are descriptive sensitivity summaries conditional on the declared perturbation set. They must not be reported as probabilities that the scientific conclusion is true.
Functions return validated AOI geometry, structured perturbation specifications and grids, perturbed geometry, assignment-comparison and stability objects, recomputed feature tables, AOI sensitivity objects, descriptive summaries, Markdown reporting text, or plots as documented by each function.
aois <- data.frame( aoi_id = c("claim", "disclosure"), xmin = c(100, 100), xmax = c(400, 400), ymin = c(100, 240), ymax = c(220, 320) ) grid <- create_aoi_perturbation_grid( dilations = c(5, 10), erosions = 5, translations_x = 5, translations_xy = c(5, -5) ) gaze <- data.frame( id = 1:5, x = c(150, 390, 405, 180, NA), y = c(150, 215, 215, 270, NA), duration = c(.1, .1, .1, .1, NA) ) result <- run_aoi_sensitivity_analysis( gaze, aois, grid, x_col = "x", y_col = "y", observation_id_col = "id", duration_col = "duration", observation_level = "fixation" ) summarise_aoi_sensitivity(result)aois <- data.frame( aoi_id = c("claim", "disclosure"), xmin = c(100, 100), xmax = c(400, 400), ymin = c(100, 240), ymax = c(220, 320) ) grid <- create_aoi_perturbation_grid( dilations = c(5, 10), erosions = 5, translations_x = 5, translations_xy = c(5, -5) ) gaze <- data.frame( id = 1:5, x = c(150, 390, 405, 180, NA), y = c(150, 215, 215, 270, NA), duration = c(.1, .1, .1, .1, NA) ) result <- run_aoi_sensitivity_analysis( gaze, aois, grid, x_col = "x", y_col = "y", observation_id_col = "id", duration_col = "duration", observation_level = "fixation" ) summarise_aoi_sensitivity(result)
Converts AOI occupancy over binned time into orthogonal-polynomial trajectory coefficients. Coefficients summarize temporal shape and are not latent psychological traits by themselves.
aoi_trajectory_features( data, person = "person_id", trial = "trial_id", time = "time_ms", aoi = "aoi", bin_ms = 100, degree = 3L, aois = NULL )aoi_trajectory_features( data, person = "person_id", trial = "trial_id", time = "time_ms", aoi = "aoi", bin_ms = 100, degree = 3L, aois = NULL )
data |
Sample-level data. |
person, trial, time, aoi
|
Column names. |
bin_ms |
Temporal bin width. |
degree |
Polynomial degree. |
aois |
Optional AOIs to encode; defaults to observed AOIs. |
An object of class "eye_aoi_trajectory", stored as a named list, with components "features", "aois", "degree", "bin_ms", "person", "trial", "caveat". It contains aOI growth-curve/trajectory features and associated metadata or diagnostics needed to interpret the result.
Map exported APIs to conceptual families
api_family_map(inventory)api_family_map(inventory)
inventory |
API inventory or character names. |
A data frame containing exported APIs to conceptual families. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compare two API lifecycle registries
api_lifecycle_diff(old, new)api_lifecycle_diff(old, new)
old |
Old registry. |
new |
New registry. |
A tabular R object containing two API lifecycle registries; rows represent analysis units and columns contain the returned quantities.
Summarise API surface by family and lifecycle status
api_surface_summary(inventory)api_surface_summary(inventory)
inventory |
Output of 'eye_api_inventory()' or compatible table. |
A data frame containing aPI surface by family and lifecycle status. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Apply a pre-flight decision to data explicitly
apply_preflight_decision( data, audit, keep_decisions = c("pass_preflight", "use_with_caution") )apply_preflight_decision( data, audit, keep_decisions = c("pass_preflight", "use_with_caution") )
data |
Original data. |
audit |
Pre-flight audit. |
keep_decisions |
Decisions to retain. |
Filtered data with an attached 'preflight_application' attribute.
Apply a synthetic corruption plan
apply_synthetic_corruption( data, plan, gaze_columns = c("gaze_x", "gaze_y"), pupil = "pupil", time = "timestamp_ms", aoi = NULL, device_column = NULL )apply_synthetic_corruption( data, plan, gaze_columns = c("gaze_x", "gaze_y"), pupil = "pupil", time = "timestamp_ms", aoi = NULL, device_column = NULL )
data |
Data frame. |
plan |
Corruption plan. |
gaze_columns |
Gaze columns receiving generic missingness/offset. |
pupil |
Pupil column. |
time |
Timestamp column. |
aoi |
Optional AOI column. |
device_column |
Optional numeric column receiving device shift. |
An R object containing a synthetic corruption plan. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Canonicalise parameter-recovery results
as_irt_recovery_results(results)as_irt_recovery_results(results)
results |
Data frame with at least 'replicate', 'parameter', 'truth', and 'estimate'; optional 'lower', 'upper', 'converged', 'scenario', 'engine', and 'failure_type' columns are retained. |
A data frame containing canonicalise parameter-recovery results. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Probabilistic AOI assignment and uncertainty propagation. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
assign_aois_probabilistic(x, aois, error_model = c("empirical", "gaussian", "ellipse"), accuracy = NULL, precision = NULL, x_col = NULL, y_col = NULL, id_cols = NULL) audit_aoi_separation(x = NULL, aois = NULL) summarise_aoi_membership(x, by = NULL) propagate_aoi_uncertainty(x, metrics = c("dwell", "ttff", "transitions", "entropy"), draws = 500, time_col = NULL, duration_col = NULL, seed = 20260807) plot_aoi_probability_map(x, ...) plot_aoi_boundary_risk(x, ...) plot_probabilistic_scanpath(x, ...) plot_fuzzy_transition_matrix(x, ...) plot_aoi_metric_uncertainty(x, ...)assign_aois_probabilistic(x, aois, error_model = c("empirical", "gaussian", "ellipse"), accuracy = NULL, precision = NULL, x_col = NULL, y_col = NULL, id_cols = NULL) audit_aoi_separation(x = NULL, aois = NULL) summarise_aoi_membership(x, by = NULL) propagate_aoi_uncertainty(x, metrics = c("dwell", "ttff", "transitions", "entropy"), draws = 500, time_col = NULL, duration_col = NULL, seed = 20260807) plot_aoi_probability_map(x, ...) plot_aoi_boundary_risk(x, ...) plot_probabilistic_scanpath(x, ...) plot_fuzzy_transition_matrix(x, ...) plot_aoi_metric_uncertainty(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
aois |
Argument controlling 'aois'; see the function usage and returned audit metadata. |
error_model |
Argument controlling 'error_model'; see the function usage and returned audit metadata. |
accuracy |
Argument controlling 'accuracy'; see the function usage and returned audit metadata. |
precision |
Argument controlling 'precision'; see the function usage and returned audit metadata. |
x_col |
Argument controlling 'x_col'; see the function usage and returned audit metadata. |
y_col |
Argument controlling 'y_col'; see the function usage and returned audit metadata. |
id_cols |
Argument controlling 'id_cols'; see the function usage and returned audit metadata. |
by |
Argument controlling 'by'; see the function usage and returned audit metadata. |
metrics |
Argument controlling 'metrics'; see the function usage and returned audit metadata. |
draws |
Argument controlling 'draws'; see the function usage and returned audit metadata. |
time_col |
Argument controlling 'time_col'; see the function usage and returned audit metadata. |
duration_col |
Argument controlling 'duration_col'; see the function usage and returned audit metadata. |
seed |
Argument controlling 'seed'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Assign features to conservative process-feature families
assign_process_feature_family( feature_names, registry = process_feature_family_registry() )assign_process_feature_family( feature_names, registry = process_feature_family_registry() )
feature_names |
Feature names. |
registry |
Feature-family registry. |
A vector or matrix containing assign features to conservative process-feature families, with shape determined by the supplied analysis units.
Identify items for descriptive 3PL/process review
audit_3pl_process_signatures( x, lower_asymptote_quantile = 0.8, fast_rt_quantile = 0.2, fast_ttff_quantile = 0.2 )audit_3pl_process_signatures( x, lower_asymptote_quantile = 0.8, fast_rt_quantile = 0.2, fast_ttff_quantile = 0.2 )
x |
An 'eye_gaze_anchored_3pl_audit'. |
lower_asymptote_quantile |
Quantile used to flag relatively large lower asymptotes. |
fast_rt_quantile |
Optional lower quantile for RT review. |
fast_ttff_quantile |
Optional lower quantile for TTFF review. |
Item-level review table. These flags are not behavioral classifications.
Audit advanced-model scientific evidence
audit_advanced_model_evidence(evidence, spec = advanced_model_evidence_spec())audit_advanced_model_evidence(evidence, spec = advanced_model_evidence_spec())
evidence |
Named list keyed by model function. Each model may contain 'recovery', 'calibration', 'misspecification', 'grouped_validation', 'engine_equivalence', 'empirical_reproduction', and 'sensitivity' objects. |
spec |
Evidence specification. |
An 'eye_advanced_evidence_audit' data frame.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
audit_benchmark_release(study = eyeprocess_benchmark_study())audit_benchmark_release(study = eyeprocess_benchmark_study())
study |
Benchmark study. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Audit bias
audit_bias(results, threshold = 0.1, by = c("scenario", "engine", "parameter"))audit_bias(results, threshold = 0.1, by = c("scenario", "engine", "parameter"))
results |
Validation or model results. |
threshold |
Decision or diagnostic threshold. |
by |
Grouping variables used when summarizing results. |
An R object containing bias. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Alias emphasizing sensitivity rather than automatic replacement
audit_biometric_imputation(...)audit_biometric_imputation(...)
... |
Additional arguments passed to the underlying method or helper. |
An R object containing alias emphasizing sensitivity rather than automatic replacement. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Aggregates trial-level signal-quality indicators by person/recording (or other supplied grouping columns) and returns review-oriented flags. No rows are automatically deleted.
audit_biometric_preflight( data, by = c("person_id"), spec = process_preflight_spec(), valid_gaze_prop = "valid_gaze_prop", valid_pupil_prop = "valid_pupil_prop", missing_gaze = "missing_gaze", missing_pupil = "missing_pupil", rt_ms = "rt_ms", blink_cluster_count = "blink_cluster_count", sampling_rate_hz = "sampling_rate_hz" )audit_biometric_preflight( data, by = c("person_id"), spec = process_preflight_spec(), valid_gaze_prop = "valid_gaze_prop", valid_pupil_prop = "valid_pupil_prop", missing_gaze = "missing_gaze", missing_pupil = "missing_pupil", rt_ms = "rt_ms", blink_cluster_count = "blink_cluster_count", sampling_rate_hz = "sampling_rate_hz" )
data |
Trial- or recording-level data. |
by |
Grouping columns, usually participant and optionally recording/session. |
spec |
A 'process_preflight_spec()' object. |
valid_gaze_prop, valid_pupil_prop
|
Column names for validity proportions. |
missing_gaze, missing_pupil
|
Column names for missingness indicators/proportions. |
rt_ms |
Response-time column. |
blink_cluster_count |
Blink-cluster count column. |
sampling_rate_hz |
Sampling-rate column. |
An 'eye_biometric_preflight' object.
Audit a candidate item bank against a seed model
audit_candidate_item_bank( object, candidate_data, difficulty_range = c(-3, 3), discrimination_min = 0.3 )audit_candidate_item_bank( object, candidate_data, difficulty_range = c(-3, 3), discrimination_min = 0.3 )
object |
Seed model. |
candidate_data |
Candidate item feature data. |
difficulty_range |
Plausible screening range for predicted difficulty. |
discrimination_min |
Minimum screening discrimination. |
An object of class "eye_candidate_item_bank_audit", stored as a named list, with components "table", "seed_model", "status", "caveat". It contains a candidate item bank against a seed model and associated metadata or diagnostics needed to interpret the result.
Audit out-of-sample incremental information from a process channel
audit_channel_incremental_information( data, fold, baseline_fitter, process_fitter, predictor, scorer, higher_is_better = TRUE )audit_channel_incremental_information( data, fold, baseline_fitter, process_fitter, predictor, scorer, higher_is_better = TRUE )
data |
Input data. |
fold |
Group/fold column. Every unique value is held out once. |
baseline_fitter |
Function fitted without the process channel. |
process_fitter |
Function fitted with the process channel. |
predictor |
Function '(fit, test)' returning predictions. |
scorer |
Function '(test, prediction)' returning a scalar score. |
higher_is_better |
Direction of the score. |
A tabular R object containing out-of-sample incremental information from a process channel; rows represent analysis units and columns contain the returned quantities.
Audit convergence and classified failures
audit_convergence(results, minimum = 0.95, by = c("scenario", "engine"))audit_convergence(results, minimum = 0.95, by = c("scenario", "engine"))
results |
Validation or model results. |
minimum |
Minimum acceptable value or threshold. |
by |
Grouping variables used when summarizing results. |
An object of class "eye_irt_convergence_audit", "data.frame", stored as a data frame, containing convergence and classified failures and associated metadata needed to interpret the result.
Audit coverage
audit_coverage( results, minimum = 0.9, maximum = 1, by = c("scenario", "engine", "parameter") )audit_coverage( results, minimum = 0.9, maximum = 1, by = c("scenario", "engine", "parameter") )
results |
Validation or model results. |
minimum |
Minimum acceptable value or threshold. |
maximum |
Maximum acceptable value or threshold. |
by |
Grouping variables used when summarizing results. |
An R object containing coverage. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Audit decision provenance and completeness
audit_decision_provenance( x, required_domains = c("sampling", "validity", "fixation", "pupil", "aoi", "model", "sensitivity", "exclusions"), required_provenance = c("data_source", "software_version", "analysis_commit") )audit_decision_provenance( x, required_domains = c("sampling", "validity", "fixation", "pupil", "aoi", "model", "sensitivity", "exclusions"), required_provenance = c("data_source", "software_version", "analysis_commit") )
x |
Manifest. |
required_domains |
Required decision domains. |
required_provenance |
Provenance keys expected under 'provenance'. |
An object of class "eye_decision_provenance_audit", stored as a named list, with components "missing_domains", "empty_domains", "missing_provenance", "complete", "manifest_hash". It contains decision provenance and completeness and associated metadata or diagnostics needed to interpret the result.
Audit distractor attention patterns
audit_distractor_attention( data, response_option = "response_option", option_gaze, chosen_suffix = NULL )audit_distractor_attention( data, response_option = "response_option", option_gaze, chosen_suffix = NULL )
data |
Input data frame or compatible tabular object. |
response_option |
Column identifying the selected response option. |
option_gaze |
Option-level gaze variables. |
chosen_suffix |
Suffix identifying the selected option indicator. |
A data frame containing distractor attention patterns. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Audit API lifecycle completeness and replacement contracts
audit_eye_api(inventory = eye_api_inventory(), registry = eye_api_lifecycle())audit_eye_api(inventory = eye_api_inventory(), registry = eye_api_lifecycle())
inventory |
API inventory. |
registry |
Lifecycle registry. |
An object of class "eye_api_audit", stored as a named list, with components "table", "unreviewed", "invalid_replacements", "invalid_canonical", "reviewed_fraction", "valid". It contains aPI lifecycle completeness and replacement contracts and associated metadata or diagnostics needed to interpret the result.
Audit a pipeline definition or completed run
audit_eye_pipeline(x)audit_eye_pipeline(x)
x |
Pipeline or pipeline run. |
An object of class "eye_pipeline_audit", stored as a named list, with components "table", "undeclared_decisions", "valid", "pipeline_hash". It contains a pipeline definition or completed run and associated metadata or diagnostics needed to interpret the result.
Audit whether a gated frontier model has a minimum evidence contract
audit_frontier_model_contract(x, evidence = list())audit_frontier_model_contract(x, evidence = list())
x |
Gated model object. |
evidence |
Named evidence objects. |
A data frame containing whether a gated frontier model has a minimum evidence contract. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Flags parameters with near-zero estimate variance, explosive dispersion, excessive missingness, or strongly correlated estimates when a covariance matrix is supplied.
audit_identifiability( results, max_missing = 0.05, max_sd_ratio = 10, correlation_matrix = NULL, max_abs_correlation = 0.995 )audit_identifiability( results, max_missing = 0.05, max_sd_ratio = 10, correlation_matrix = NULL, max_abs_correlation = 0.995 )
results |
Validation or model results. |
max_missing |
Maximum acceptable missingness. |
max_sd_ratio |
Maximum acceptable standard-deviation ratio. |
correlation_matrix |
Optional parameter-correlation matrix. |
max_abs_correlation |
Maximum acceptable absolute parameter correlation. |
An object of class "eye_irt_identifiability_audit", "data.frame", stored as a data frame, containing empirical identifiability from replicate estimates and associated metadata needed to interpret the result.
Audit interval width
audit_interval_width( results, maximum = Inf, by = c("scenario", "engine", "parameter") )audit_interval_width( results, maximum = Inf, by = c("scenario", "engine", "parameter") )
results |
Validation or model results. |
maximum |
Maximum acceptable value or threshold. |
by |
Grouping variables used when summarizing results. |
An R object containing interval width. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Audit item response-function shape departures
audit_irf_shape( comparison, mean_absolute_threshold = 0.05, max_absolute_threshold = 0.15 )audit_irf_shape( comparison, mean_absolute_threshold = 0.05, max_absolute_threshold = 0.15 )
comparison |
Value supplied to 'comparison'; see Details for its model-specific role. |
mean_absolute_threshold |
Threshold for mean absolute IRF departure. |
max_absolute_threshold |
Threshold for maximum absolute IRF departure. |
A tabular R object containing item response-function shape departures; rows represent analysis units and columns contain the returned quantities.
Run stepwise Rasch item-reduction as a sensitivity analysis
audit_item_reduction_sensitivity( erm_model, criterion = list("itemfit"), alpha = 0.05, maxstep = 5L )audit_item_reduction_sensitivity( erm_model, criterion = list("itemfit"), alpha = 0.05, maxstep = 5L )
erm_model |
Fitted eRm model. |
criterion |
Criterion list passed to 'eRm::stepwiseIt()'. |
alpha |
Significance threshold. |
maxstep |
Maximum elimination steps. |
An object of class "eye_item_reduction_sensitivity", stored as a named list, with components "model", "eliminated_items", "alpha", "maxstep", "status", "caveat". It contains stepwise Rasch item-reduction as a sensitivity analysis and associated metadata or diagnostics needed to interpret the result.
Audit the empirical latent-trait distribution
audit_latent_distribution(theta, tail_z = 3)audit_latent_distribution(theta, tail_z = 3)
theta |
Numeric latent-trait draws/estimates. |
tail_z |
Absolute standardized threshold used for tail-rate diagnostics. |
A data frame containing the empirical latent-trait distribution. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Summarise measurement transportability across held-out groups
audit_measurement_transportability( validation, metric, higher_is_better = TRUE, max_range = NULL, minimum = NULL, maximum = NULL )audit_measurement_transportability( validation, metric, higher_is_better = TRUE, max_range = NULL, minimum = NULL, maximum = NULL )
validation |
Validation results or validation specification. |
metric |
Metric to calculate or audit. |
higher_is_better |
Whether larger metric values indicate better performance. |
max_range |
Maximum allowed range across held-out groups. |
minimum |
Minimum acceptable value or threshold. |
maximum |
Maximum acceptable value or threshold. |
An object of class "eye_measurement_transportability_audit", "data.frame", stored as a data frame, containing measurement transportability across held-out groups and associated metadata needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
audit_model_promotion(evidence, spec = model_promotion_spec())audit_model_promotion(evidence, spec = model_promotion_spec())
evidence |
Named list by model family. Each family may contain |
spec |
Promotion specification. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
This is a structural pre-flight screen, not a proof of statistical identifiability.
audit_multimodal_identifiability(x, min_person = 30L, min_item = 5L)audit_multimodal_identifiability(x, min_person = 30L, min_item = 5L)
x |
An 'eye_multimodal_measurement'. |
min_person, min_item
|
Minimum structural counts. |
An 'eye_multimodal_identifiability_audit'.
This is a conservative pre-fit audit. It does not prove global identifiability. It verifies support, design connectivity, channel coverage, response variation, and the explicit scale constraints used by the reference likelihoods.
audit_multimodal_m2_identifiability( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", model = c("M2", "M1", "M0"), min_persons = 20L, min_items = 5L )audit_multimodal_m2_identifiability( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", model = c("M2", "M1", "M0"), min_persons = 20L, min_items = 5L )
x |
Data frame or compatible M2 object. |
person, item, response, rt, gaze
|
Column names. |
model |
'"M0"', '"M1"', or '"M2"'. |
min_persons, min_items
|
Conservative design thresholds. |
An 'eye_multimodal_m2_identifiability' object.
Performs a conservative pre-fit audit for the four-channel response, RT, fixation-count and pupil reference model. The audit checks design connectivity, channel coverage and variation, pupil nuisance availability, blink/interpolation burden, device/session representation and explicit M3 identification constraints. It is a support screen, not proof of global identifiability or construct validity.
audit_multimodal_m3_identifiability( x, pupil_scale = c("z", "raw"), min_persons = 20L, min_items = 5L, max_pupil_missing = 0.5, max_blink_rate = 0.3, max_interpolation_rate = 0.3 )audit_multimodal_m3_identifiability( x, pupil_scale = c("z", "raw"), min_persons = 20L, min_items = 5L, max_pupil_missing = 0.5, max_blink_rate = 0.3, max_interpolation_rate = 0.3 )
x |
M3-compatible data, simulation, fit, or measurement object. |
pupil_scale |
Pupil transformation used by the reference likelihood. |
min_persons, min_items
|
Conservative design thresholds. |
max_pupil_missing, max_blink_rate, max_interpolation_rate
|
Warning thresholds. |
An 'eye_multimodal_m3_identifiability' object.
Evaluates sequence information, channel support, state occupancy, assignment uncertainty, state separation, transition degeneracy, and HMC diagnostics. The result is deliberately multi-criterion; M4 does not collapse validity to a single undocumented boolean.
audit_multimodal_m4_identifiability( x, spec = NULL, include_posterior = TRUE, rhat_max = 1.05, ess_min = 100, ebfmi_min = 0.3, occupancy_min = 0.03, entropy_fraction_review = 0.8 )audit_multimodal_m4_identifiability( x, spec = NULL, include_posterior = TRUE, rhat_max = 1.05, ess_min = 100, ebfmi_min = 0.3, occupancy_min = 0.03, entropy_fraction_review = 0.8 )
x |
Data, M4 simulation, M4 fit, or internal prepared M4 data. |
spec |
Optional M4 specification when 'x' is not a fit. |
include_posterior |
Whether to evaluate posterior criteria for a fit. |
rhat_max, ess_min, ebfmi_min
|
Sampler thresholds. |
occupancy_min |
Minimum mean posterior state occupancy for non-null K. |
entropy_fraction_review |
Review threshold relative to maximum entropy. |
An 'eye_multimodal_m4_identifiability' with machine-readable checks.
Audit a multimodal measurement object
audit_multimodal_measurement(x)audit_multimodal_measurement(x)
x |
An 'eye_multimodal_measurement'. |
An 'eye_multimodal_audit'.
Alias emphasizing data-quality interpretation of process anomaly auditing
audit_multivariate_process_quality(...)audit_multivariate_process_quality(...)
... |
Additional arguments passed to the underlying method or helper. |
An R object containing alias emphasizing data-quality interpretation of process anomaly auditing. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Run nonparametric Rasch diagnostics with eRm
audit_nonparametric_rasch( response_matrix, methods = c("T1", "T10"), n = 100L, splitcr = "median", seed = 321 )audit_nonparametric_rasch( response_matrix, methods = c("T1", "T10"), n = 100L, splitcr = "median", seed = 321 )
response_matrix |
Dichotomous response matrix. |
methods |
eRm NPtest methods, e.g. T1 and T10. |
n |
Number of sampled matrices. |
splitcr |
Split criterion for tests that use one. |
seed |
Seed. |
An object of class "eye_nonparametric_rasch_audit", stored as a named list, with components "tests", "status", "n", "splitcr", "caveat". It contains nonparametric Rasch diagnostics with eRm and associated metadata or diagnostics needed to interpret the result.
Creates a transparent presentation-review score from response time/dwell, revisit/entropy, pupil-effort proxies, and gaze quality. It is an experimental design/fairness audit only.
audit_presentation_accessibility( data, person = "person_id", rt = "rt_ms", dwell = "dwell_ms", revisits = "revisits", entropy = "aoi_entropy", pupil = "pupil_peak", gaze_validity = "valid_gaze_prop", review_quantile = 0.9 )audit_presentation_accessibility( data, person = "person_id", rt = "rt_ms", dwell = "dwell_ms", revisits = "revisits", entropy = "aoi_entropy", pupil = "pupil_peak", gaze_validity = "valid_gaze_prop", review_quantile = 0.9 )
data |
Process data. |
person |
Person identifier. |
rt, dwell, revisits, entropy, pupil, gaze_validity
|
Optional column names. |
review_quantile |
Quantile for a presentation-review flag. |
An object of class "eye_presentation_accessibility", stored as a named list, with components "table", "threshold", "review_quantile", "status", "caveat". It contains presentation/accessibility sensitivity without clinical inference and associated metadata or diagnostics needed to interpret the result.
Audit DIF before and after process-data adjustment
audit_process_adjusted_dif( data, response = "response", ability, group, item = "item_id", process_features, person = "participant_id" )audit_process_adjusted_dif( data, response = "response", ability, group, item = "item_id", process_features, person = "participant_id" )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
ability |
Value supplied to 'ability'; see Details for its model-specific role. |
group |
Value supplied to 'group'; see Details for its model-specific role. |
item |
Item identifier, name, or item column. |
process_features |
Names of process-derived features. |
person |
Person or participant identifier column. |
An object of class "eye_process_adjusted_dif", stored as a named list, with components "unadjusted_model", "adjusted_model", "coefficients", "surrogate", "note". It contains dIF before and after process-data adjustment and associated metadata or diagnostics needed to interpret the result.
Computes a regularized Mahalanobis distance over selected person-level process metrics. Flags indicate review needs only; they are not cheating, identity, diagnosis, or intent classifications.
audit_process_anomalies( data, person = "person_id", metrics = NULL, alpha = 0.975, aggregate = TRUE, ridge = 1e-06 )audit_process_anomalies( data, person = "person_id", metrics = NULL, alpha = 0.975, aggregate = TRUE, ridge = 1e-06 )
data |
Data frame. |
person |
Person identifier column. |
metrics |
Numeric process metrics. If omitted, usable numeric columns are selected. |
alpha |
Chi-square review quantile. |
aggregate |
If TRUE, aggregate metrics to person level before auditing. |
ridge |
Diagonal covariance regularization. |
An object of class "eye_process_anomaly_audit", stored as a named list, with components "table", "metrics", "alpha", "threshold", "center", "covariance", "caveat". It contains multivariate process/data-quality anomalies and associated metadata or diagnostics needed to interpret the result.
Audit post-deployment psychometric and biometric drift
audit_process_drift( data, item = "item_id", batch = "deployment_batch", metrics = c("irt_difficulty", "irt_discrimination", "rt_ms", "dwell_ms", "pupil_bc", "valid_gaze_prop", "screen_luminance"), spec = process_drift_spec(), reference_batch = NULL, aggregate_fun = .ep08_mean )audit_process_drift( data, item = "item_id", batch = "deployment_batch", metrics = c("irt_difficulty", "irt_discrimination", "rt_ms", "dwell_ms", "pupil_bc", "valid_gaze_prop", "screen_luminance"), spec = process_drift_spec(), reference_batch = NULL, aggregate_fun = .ep08_mean )
data |
Item-by-batch or trial-level deployment data. |
item |
Item identifier column. |
batch |
Ordered deployment batch/date column. |
metrics |
Numeric metrics to monitor. |
spec |
Drift specification. |
reference_batch |
Optional reference batch value(s). |
aggregate_fun |
Aggregation function used when multiple rows occur within item x batch. |
An object of class "eye_process_drift_audit", stored as a named list, with components "table", "trajectories", "item", "batch", "metrics", "spec", "reference_batch", "flag_columns", "caveat". It contains post-deployment psychometric and biometric drift and associated metadata or diagnostics needed to interpret the result.
Audit external/structural validity of process traits
audit_process_external_validity( data, criterion, predictors, baseline_predictors = NULL )audit_process_external_validity( data, criterion, predictors, baseline_predictors = NULL )
data |
Person-level data containing a criterion and process predictors. |
criterion |
External criterion column. |
predictors |
Process predictors. |
baseline_predictors |
Optional baseline predictors for incremental validity. |
An object of class "eye_process_external_validity", stored as a named list, with components "full_model", "baseline_model", "comparison", "associations", "criterion", "predictors", "baseline_predictors", "data", "incremental_r2", "status", "caveat". It contains external/structural validity of process traits and associated metadata or diagnostics needed to interpret the result.
Computes pairwise residual correlations (a Q3-style diagnostic) across item or option columns and, when supplied, analogous process-residual correlations. This is a diagnostic for local dependence, not a formal test with universal cutoffs.
audit_process_local_dependence( response_residuals, process_residuals = NULL, threshold = 0.2 )audit_process_local_dependence( response_residuals, process_residuals = NULL, threshold = 0.2 )
response_residuals |
Person-by-item/option residual matrix. |
process_residuals |
Optional aligned process-residual matrix. |
threshold |
Absolute correlation threshold used only for flagging. |
An object of class "eye_process_local_dependence_audit", stored as a named list, with components "pairs", "threshold", "max_absolute_response", "note". It contains inter-option/process local dependence and associated metadata or diagnostics needed to interpret the result.
Audit process measurement invariance across facets
audit_process_measurement_invariance( object, channel = c("process", "response"), relative_sd_threshold = 0.25 )audit_process_measurement_invariance( object, channel = c("process", "response"), relative_sd_threshold = 0.25 )
object |
A fitted eyeprocess model or audit object. |
channel |
Measurement channel to inspect. |
relative_sd_threshold |
Value supplied to 'relative_sd_threshold'; see Details for its model-specific role. |
An object of class "eye_process_measurement_invariance", stored as a named list, with components "pass", "threshold", "components", "channel", "note". It contains process measurement invariance across facets and associated metadata or diagnostics needed to interpret the result.
Audit sensitivity of process summaries to temporal window choices
audit_process_window_sensitivity( data, widths_ms = c(250, 500, 1000, 1500), steps_ms = c(100, 250, 500), metric = "pupil_mean", grid = TRUE, ... )audit_process_window_sensitivity( data, widths_ms = c(250, 500, 1000, 1500), steps_ms = c(100, 250, 500), metric = "pupil_mean", grid = TRUE, ... )
data |
Sample-level data. |
widths_ms |
Window widths. |
steps_ms |
Step widths; recycled or crossed depending on 'grid'. |
metric |
Extracted process metric to compare. |
grid |
If TRUE, evaluate all width-step combinations. |
... |
Passed to 'extract_process_windows()'. |
An object of class "eye_process_window_sensitivity", stored as a named list, with components "table", "metric", "settings", "caveat". It contains sensitivity of process summaries to temporal window choices and associated metadata or diagnostics needed to interpret the result.
Audit within-person pupil fatigue/trial-order drift
audit_pupil_fatigue_drift( data, pupil = "pupil_peak", trial_order = "trial_sequence", person = "person_id", luminance = NULL, difficulty = NULL, engine = c("auto", "plm", "lm_fixed_effects") )audit_pupil_fatigue_drift( data, pupil = "pupil_peak", trial_order = "trial_sequence", person = "person_id", luminance = NULL, difficulty = NULL, engine = c("auto", "plm", "lm_fixed_effects") )
data |
Trial-level data. |
pupil, trial_order, person
|
Required columns. |
luminance, difficulty
|
Optional covariates. |
engine |
'auto', 'plm', or 'lm_fixed_effects'. |
An object of class "eye_pupil_fatigue_drift", stored as a named list, with components "model", "coefficients", "data", "engine", "status", "caveat". It contains within-person pupil fatigue/trial-order drift and associated metadata or diagnostics needed to interpret the result.
Audit stability of pupil frequency features across window lengths
audit_pupil_frequency_stability( data, windows_ms = c(500, 1000, 2000), by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", sampling_rate_hz = 60 )audit_pupil_frequency_stability( data, windows_ms = c(500, 1000, 2000), by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", sampling_rate_hz = 60 )
data |
Sample-level data. |
windows_ms |
Window lengths to evaluate. |
by, time, pupil, sampling_rate_hz
|
Passed through to feature construction. |
An object of class "eye_pupil_frequency_stability", stored as a named list, with components "table", "windows_ms", "caveat". It contains stability of pupil frequency features across window lengths and associated metadata or diagnostics needed to interpret the result.
Audit declared order of pupil preprocessing steps
audit_pupil_preprocessing_order( steps, cleaning_patterns = c("blink", "missing", "interpol", "artifact", "smooth", "filter"), baseline_pattern = "baseline" )audit_pupil_preprocessing_order( steps, cleaning_patterns = c("blink", "missing", "interpol", "artifact", "smooth", "filter"), baseline_pattern = "baseline" )
steps |
Character vector in execution order. |
cleaning_patterns |
Patterns considered cleaning/preprocessing. |
baseline_pattern |
Pattern identifying baseline correction. |
A named list with components "steps", "baseline_positions", "cleaning_positions", "cleaning_after_baseline", "status", "caveat", containing declared order of pupil preprocessing steps and associated metadata or diagnostics.
Audit rmse
audit_rmse(results, threshold = 0.3, by = c("scenario", "engine", "parameter"))audit_rmse(results, threshold = 0.3, by = c("scenario", "engine", "parameter"))
results |
Validation or model results. |
threshold |
Decision or diagnostic threshold. |
by |
Grouping variables used when summarizing results. |
An R object containing rmse. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
audit_roundtrip_loss(source, roundtrip, tables = canonical_table_names(), tolerance = 1e-8)audit_roundtrip_loss(source, roundtrip, tables = canonical_table_names(), tolerance = 1e-8)
source |
Source and re-imported 'eye_dataset' objects. |
roundtrip |
Source and re-imported 'eye_dataset' objects. |
tables |
Canonical tables to compare. |
tolerance |
Numeric tolerance. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Audit sampling irregularity
audit_sampling_irregularity( data, time = "timestamp_ms", unit = c("ms", "s", "us"), by = NULL, cv_threshold = 0.05 )audit_sampling_irregularity( data, time = "timestamp_ms", unit = c("ms", "s", "us"), by = NULL, cv_threshold = 0.05 )
data |
Sample data. |
time |
Timestamp column. |
unit |
Timestamp unit. |
by |
Optional grouping columns. |
cv_threshold |
Review threshold for interval coefficient of variation. |
An object of class "eye_sampling_irregularity_audit", stored as a named list, with components "table", "cv_threshold", "caveat". It contains sampling irregularity and associated metadata or diagnostics needed to interpret the result.
Uses binned chi-square diagnostics as a coarse screening diagnostic and also reports the mean/variance of normalized ranks. It is not a replacement for rank-histogram inspection.
audit_sbc(x, bins = 10L, alpha = 0.01)audit_sbc(x, bins = 10L, alpha = 0.01)
x |
Object to print, plot, summarize, or audit. |
bins |
Number of bins used by the diagnostic. |
alpha |
Significance or tail-probability level. |
An object of class "eye_sbc_audit", "data.frame", stored as a data frame, containing sBC rank uniformity and associated metadata needed to interpret the result.
Summarize a signal-filter audit
audit_signal_filter(x)audit_signal_filter(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing a signal-filter audit. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Leakage here means information becoming available after the declared outcome boundary. It is a data/analysis property and is not an allegation of misconduct.
audit_temporal_leakage(provenance, allow_equal = TRUE, tolerance = 0)audit_temporal_leakage(provenance, allow_equal = TRUE, tolerance = 0)
provenance |
Output of [process_feature_time_provenance()] or compatible table. |
allow_equal |
Whether features available exactly at outcome time are allowed. |
tolerance |
Numeric tolerance in the provenance time unit. |
A named list with components "status", "n_features", "n_flagged", "flagged_fraction", "detail", "interpretation", containing temporal leakage in a feature provenance table and associated metadata or diagnostics.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
audit_validation_completion(x, thresholds = validation_thresholds(), empirical_reproduction = NULL)audit_validation_completion(x, thresholds = validation_thresholds(), empirical_reproduction = NULL)
x |
Validation collection. |
thresholds |
Validation thresholds. |
empirical_reproduction |
Optional empirical-reproduction evidence required when configured in 'thresholds'. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
audit_vendor_field_coverage(semantics, required_fields)audit_vendor_field_coverage(semantics, required_fields)
semantics |
Semantics registry or corpus path. |
required_fields |
Data frame with canonical table/field pairs. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Audit a multi-vendor validation corpus
audit_vendor_validation(x, spec = vendor_validation_spec())audit_vendor_validation(x, spec = vendor_validation_spec())
x |
An 'eye_corpus_validation' object or its summary data frame. |
spec |
Validation specification. |
An 'eye_vendor_validation' data frame.
Audit visual-context dependence
audit_visual_context_dependence(x)audit_visual_context_dependence(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing visual-context dependence. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Extract compact Bayesian process-model diagnostic flags
bayesian_process_diagnostic_flags( x, rhat_threshold = 1.01, ess_threshold = 400 )bayesian_process_diagnostic_flags( x, rhat_threshold = 1.01, ess_threshold = 400 )
x |
An 'eye_bayesian_process_dashboard'. |
rhat_threshold |
Review threshold for R-hat. |
ess_threshold |
Review threshold for bulk/tail effective sample size. |
An R object containing compact Bayesian process-model diagnostic flags. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Collects model availability, optional approximate leave-one-out summaries, posterior diagnostics, and (only when explicitly requested) a Bayes-factor comparison. The function does not treat any one diagnostic as proof of the substantive process interpretation.
bayesian_process_diagnostics_dashboard( ..., model_names = NULL, compute_loo = TRUE, compute_bayes_factor = FALSE, posterior_summary = TRUE )bayesian_process_diagnostics_dashboard( ..., model_names = NULL, compute_loo = TRUE, compute_bayes_factor = FALSE, posterior_summary = TRUE )
... |
Fitted 'brmsfit' objects. |
model_names |
Optional model labels. |
compute_loo |
Whether to compute approximate leave-one-out diagnostics. |
compute_bayes_factor |
Whether to attempt a Bayes factor. This requires exactly two suitable brms models and typically models fitted with 'save_pars = save_pars(all = TRUE)'. |
posterior_summary |
Whether to collect posterior convergence summaries when package 'posterior' is available. |
An 'eye_bayesian_process_dashboard' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
benchmark_expected_outputs(study = eyeprocess_benchmark_study())benchmark_expected_outputs(study = eyeprocess_benchmark_study())
study |
Benchmark object or path. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
benchmark_eye_storage(x, formats = c("rds", "csv", "parquet"), partition_by = c("participant_id", "recording_id"), repetitions = 3L, directory = tempdir())benchmark_eye_storage(x, formats = c("rds", "csv", "parquet"), partition_by = c("participant_id", "recording_id"), repetitions = 3L, directory = tempdir())
x |
Named list of tables or eye dataset. |
formats |
Formats to benchmark. |
partition_by |
Partition columns. |
repetitions |
Repetitions. |
directory |
Parent temporary directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Benchmark an eyeprocess operation
benchmark_eyeprocess(expr, iterations = 5L, label = "operation")benchmark_eyeprocess(expr, iterations = 5L, label = "operation")
expr |
Function with no arguments to benchmark. |
iterations |
Number of iterations. |
label |
Benchmark label. |
An 'eye_benchmark' data frame.
Memory estimate for an R object or generated problem size
benchmark_memory_estimate(x, generator = NULL)benchmark_memory_estimate(x, generator = NULL)
x |
Object, or numeric n when 'generator' is supplied. |
generator |
Optional function taking n. |
A data frame containing memory estimate for an R object or generated problem size. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Estimate scaling exponent from benchmark results
benchmark_scaling_curve(x)benchmark_scaling_curve(x)
x |
Benchmark result. |
A data frame containing scaling exponent from benchmark results. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Bind compatible process-window objects
bind_process_windows(...)bind_process_windows(...)
... |
Additional arguments passed to the underlying method or helper. |
An object of class "eye_process_windows", stored as a named list, with components "data", "spec", "source_n", "status". It contains bind compatible process-window objects and associated metadata or diagnostics needed to interpret the result.
Run biometric-feature imputation as a sensitivity analysis
biometric_imputation_sensitivity( data, variables, methods = c("mice", "missForest"), m = 3L, maxit = 3L, seed = 521 )biometric_imputation_sensitivity( data, variables, methods = c("mice", "missForest"), m = 3L, maxit = 3L, seed = 521 )
data |
Data containing biometric/process features. |
variables |
Variables to impute. |
methods |
Any of 'mice' and 'missForest'. |
m |
Number of MICE imputations. |
maxit |
Iteration count. |
seed |
Seed. |
An 'eye_biometric_imputation_sensitivity' object. Complete-case analysis is not replaced automatically.
Bootstrap ICC reliability by resampling participants
bootstrap_process_reliability( data, person, session, measure, replications = 500L, seed = 1L )bootstrap_process_reliability( data, person, session, measure, replications = 500L, seed = 1L )
data |
Long data. |
person, session, measure
|
Column names. |
replications |
Bootstrap replications. |
seed |
Seed. |
A data frame containing bootstrap ICC reliability by resampling participants. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
build_compatibility_matrix(x, required_vendors = c("gazepoint", "tobii", "pupillabs", "eyelink", "smi"), min_empirical_cases = 2L)build_compatibility_matrix(x, required_vendors = c("gazepoint", "tobii", "pupillabs", "eyelink", "smi"), min_empirical_cases = 2L)
x |
Corpus path or registry data frame. |
required_vendors |
Required vendors. |
min_empirical_cases |
Minimum independent empirical cases per vendor. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Evidence and decision provenance graphs. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
build_evidence_graph(raw_data, transformations = NULL, metrics = NULL, models = NULL, diagnostics = NULL, decisions = NULL, edges = NULL) trace_item_decision(graph, item_id) compare_decision_provenance(graph_a, graph_b) audit_evidence_dependencies(graph) plot_evidence_graph(x, ...) plot_item_decision_path(x, ...) plot_metric_dependency_graph(x, ...) plot_model_decision_impact(x, ...)build_evidence_graph(raw_data, transformations = NULL, metrics = NULL, models = NULL, diagnostics = NULL, decisions = NULL, edges = NULL) trace_item_decision(graph, item_id) compare_decision_provenance(graph_a, graph_b) audit_evidence_dependencies(graph) plot_evidence_graph(x, ...) plot_item_decision_path(x, ...) plot_metric_dependency_graph(x, ...) plot_model_decision_impact(x, ...)
raw_data |
Argument controlling 'raw_data'; see the function usage and returned audit metadata. |
transformations |
Argument controlling 'transformations'; see the function usage and returned audit metadata. |
metrics |
Argument controlling 'metrics'; see the function usage and returned audit metadata. |
models |
Argument controlling 'models'; see the function usage and returned audit metadata. |
diagnostics |
Argument controlling 'diagnostics'; see the function usage and returned audit metadata. |
decisions |
Argument controlling 'decisions'; see the function usage and returned audit metadata. |
edges |
Argument controlling 'edges'; see the function usage and returned audit metadata. |
graph |
Argument controlling 'graph'; see the function usage and returned audit metadata. |
item_id |
Argument controlling 'item_id'; see the function usage and returned audit metadata. |
graph_a |
Argument controlling 'graph_a'; see the function usage and returned audit metadata. |
graph_b |
Argument controlling 'graph_b'; see the function usage and returned audit metadata. |
x |
Input object or data structure appropriate for the selected analysis. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Converts contiguous Gazepoint media runs into explicit trial intervals. This provides stable person-by-item-by-trial keys even when the original export contains no behavioural response file.
build_gazepoint_media_trials(x, item_map = NULL, overwrite = TRUE)build_gazepoint_media_trials(x, item_map = NULL, overwrite = TRUE)
x |
An 'eye_dataset' imported from Gazepoint. |
item_map |
Optional data frame or CSV path containing 'stimulus_id', 'item_id', and optionally 'condition_id'. |
overwrite |
Replace existing trial intervals. |
The updated 'eye_dataset'.
Calibration drift profile across sessions/batches
calibration_drift_profile( data, by, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y" )calibration_drift_profile( data, by, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y" )
data |
Validation-target data. |
by |
Ordered batch/session column. |
gaze_x, gaze_y
|
Recorded gaze-coordinate columns. |
target_x, target_y
|
Known target-coordinate columns. |
An object of class "eye_calibration_drift_profile", stored as a named list, with components "table", "by", "caveat". It contains calibration drift profile across sessions/batches and associated metadata or diagnostics needed to interpret the result.
Build an empirical bivariate calibration-error model
calibration_error_model( data, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y" )calibration_error_model( data, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y" )
data |
Validation-target data. |
gaze_x, gaze_y
|
Recorded gaze-coordinate columns. |
target_x, target_y
|
Known target-coordinate columns. |
An object of class "eye_calibration_error_model", stored as a named list, with components "mean_error", "covariance", "errors", "n", "metrics", "coordinate_units", "status", "caveat". It contains an empirical bivariate calibration-error model and associated metadata or diagnostics needed to interpret the result.
Sensitivity grid for deterministic calibration offsets
calibration_sensitivity_grid( offset_x = c(-0.02, 0, 0.02), offset_y = c(-0.02, 0, 0.02) )calibration_sensitivity_grid( offset_x = c(-0.02, 0, 0.02), offset_y = c(-0.02, 0, 0.02) )
offset_x, offset_y
|
Candidate offsets in coordinate units. |
A tabular R object containing sensitivity grid for deterministic calibration offsets; rows represent analysis units and columns contain the returned quantities.
Audit transfer of calibration across devices/sessions/sites
calibration_transfer_audit(data, group, observed, predicted)calibration_transfer_audit(data, group, observed, predicted)
data |
Data containing group, observed and predicted values. |
group |
Grouping column. |
observed |
Observed binary/numeric outcome column. |
predicted |
Predicted probability/numeric score column. |
An object of class "eye_calibration_transfer_audit", "data.frame", stored as a data frame, containing transfer of calibration across devices/sessions/sites and associated metadata needed to interpret the result.
Canonical API mapping
canonical_eye_api(registry = eye_api_lifecycle())canonical_eye_api(registry = eye_api_lifecycle())
registry |
Lifecycle registry. |
A tabular R object containing canonical API mapping; rows represent analysis units and columns contain the returned quantities.
Classify item missingness using exposure and response evidence
classify_item_missingness( data, response = "response", reached = "reached", inspected = NULL, started = NULL )classify_item_missingness( data, response = "response", reached = "reached", inspected = NULL, started = NULL )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
reached |
Indicator that the item was reached. |
inspected |
Indicator that the item or response area was inspected. |
started |
Indicator that responding was initiated. |
An object of class "factor", stored as an R object, containing classify item missingness using exposure and response evidence and associated metadata needed to interpret the result.
Collect a disk-backed eye dataset
collect_eye_storage(x, tables = NULL)collect_eye_storage(x, tables = NULL)
x |
An 'eye_storage' handle or storage path. |
tables |
Optional subset of canonical tables. |
An 'eye_dataset'.
Collect validation evidence into a common bundle
collect_validation_evidence(..., model_name = NULL, notes = NULL)collect_validation_evidence(..., model_name = NULL, notes = NULL)
... |
Named validation objects. |
model_name |
Optional model/workflow label. |
notes |
Optional free-text notes. |
An 'eye_validation_bundle' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
collect_validation_jobs(path, plan = NULL, strict = TRUE)collect_validation_jobs(path, plan = NULL, strict = TRUE)
path |
Manifest/output directory or character vector of directories. |
plan |
Optional validation plan. |
strict |
Fail when duplicate job identifiers disagree. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare explicit AOI assignment methods
compare_aoi_methods( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )compare_aoi_methods( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )
data |
Data. |
methods |
Named methods/specifications. |
analysis_fun |
Function '(data, method, specification)'. |
extract_fun |
Result extractor. |
An object of class "eye_process_sensitivity", stored as a named list, with components "grid", "results", "failures", "warnings", "grid_hash", "created_at", "status", "caveat". It contains explicit AOI assignment methods and associated metadata or diagnostics needed to interpret the result.
Compare AOI trajectory feature objects
compare_aoi_trajectories(...)compare_aoi_trajectories(...)
... |
Additional arguments passed to the underlying method or helper. |
A data frame containing aOI trajectory feature objects. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compare Bayesian process models by LOO or Bayes factor
compare_bayesian_process_models(..., method = c("loo", "bayes_factor"))compare_bayesian_process_models(..., method = c("loo", "bayes_factor"))
... |
Fitted brms models. |
method |
'loo' or 'bayes_factor'. |
An R object containing bayesian process models by LOO or Bayes factor. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Compare two research decision manifests
compare_decision_manifests(old, new)compare_decision_manifests(old, new)
old |
Earlier manifest. |
new |
Later manifest. |
An object of class "eye_decision_manifest_diff", "data.frame", stored as a data frame, containing two research decision manifests and associated metadata needed to interpret the result.
Compare two deployment batches descriptively
compare_deployment_batches( data, batch = "deployment_batch", batch_a, batch_b, metrics = NULL, item = "item_id" )compare_deployment_batches( data, batch = "deployment_batch", batch_a, batch_b, metrics = NULL, item = "item_id" )
data |
Deployment data. |
batch |
Batch column. |
batch_a, batch_b
|
Values to compare. |
metrics |
Metrics to compare. |
item |
Optional item identifier for item-matched differences. |
A data frame containing two deployment batches descriptively. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_diffusion_accuracy_rt(object)compare_diffusion_accuracy_rt(object)
object |
Diffusion fit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_dynamic_transition_models(..., criterion = c("AIC", "BIC", "log_score", "accuracy"))compare_dynamic_transition_models(..., criterion = c("AIC", "BIC", "log_score", "accuracy"))
... |
Named fitted dynamic IRTree models. |
criterion |
AIC, BIC, log score, or classification accuracy. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_engine_adapters(...)compare_engine_adapters(...)
... |
Adapter results or a list. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare explicit fixation-detection methods
compare_fixation_methods( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )compare_fixation_methods( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )
data |
Data. |
methods |
Named methods/specifications. |
analysis_fun |
Function '(data, method, specification)'. |
extract_fun |
Result extractor. |
An object of class "eye_process_sensitivity", stored as a named list, with components "grid", "results", "failures", "warnings", "grid_hash", "created_at", "status", "caveat". It contains explicit fixation-detection methods and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_functional_scalar_models(x, scalar_features = c("pupil_peak", "pupil_auc", "pupil_mean"), criterion = c("AIC", "log_loss"), folds = 5L, seed = 1L)compare_functional_scalar_models(x, scalar_features = c("pupil_peak", "pupil_auc", "pupil_mean"), criterion = c("AIC", "log_loss"), folds = 5L, seed = 1L)
x |
Functional pupil fit or prepared data. |
scalar_features |
Scalar feature names. |
criterion |
AIC or cross-validated log loss. |
folds |
Grouped folds for cross-validation. |
seed |
Random seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare hard and probabilistic AOI assignments
compare_hard_probabilistic_aoi( data, aois, probabilistic, x = "gaze_x", y = "gaze_y" )compare_hard_probabilistic_aoi( data, aois, probabilistic, x = "gaze_x", y = "gaze_y" )
data |
Gaze samples. |
aois |
Rectangular AOIs. |
probabilistic |
Probabilistic assignment object. |
x, y
|
Gaze-coordinate columns. |
An R object containing hard and probabilistic AOI assignments. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Uses supplied scoring functions when models expose different engines. The default extracts AIC/BIC/logLik when available and never treats in-sample fit as sufficient evidence for model promotion.
compare_irt_models(..., names = NULL)compare_irt_models(..., names = NULL)
... |
Additional arguments passed to the selected model, engine, or method. |
names |
Value supplied to 'names'; see Details for its model-specific role. |
An R object containing multimodal IRT model objects. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Compares Gaussian, location/scale Student-t, and two-normal-mixture reference densities by AIC/BIC. This is a stress-test diagnostic; it does not change the latent distribution inside an already fitted IRT model.
compare_latent_distribution_models(theta)compare_latent_distribution_models(theta)
theta |
Numeric latent-trait draws/estimates. |
An object of class "eye_latent_distribution_comparison", stored as a named list, with components "comparison", "audit", "student_t", "mixture", "status". It contains simple latent-distribution reference models and associated metadata or diagnostics needed to interpret the result.
Compare equivalent model engines
compare_model_engines( data, engines, extractors, reference = names(engines)[1L], tolerance = 0.05 )compare_model_engines( data, engines, extractors, reference = names(engines)[1L], tolerance = 0.05 )
data |
Model-ready data. |
engines |
Named list of fitting functions receiving 'data'. |
extractors |
Named list of extractor functions or one shared extractor. |
reference |
Optional reference engine name. |
tolerance |
Maximum absolute estimate difference for equivalence. |
An 'eye_engine_comparison' object.
Compare conventional logistic and flexible IRF shapes
compare_parametric_nonparametric_irf( response_matrix, gpirt_object = NULL, theta_grid = seq(-4, 4, length.out = 101) )compare_parametric_nonparametric_irf( response_matrix, gpirt_object = NULL, theta_grid = seq(-4, 4, length.out = 101) )
response_matrix |
Person-by-item response matrix. |
gpirt_object |
Value supplied to 'gpirt_object'; see Details for its model-specific role. |
theta_grid |
Grid of latent-trait values used for evaluation. |
An object of class "eye_irf_comparison", "data.frame", stored as a data frame, containing conventional logistic and flexible IRF shapes and associated metadata needed to interpret the result.
Compare outcomes across presentation variants
compare_presentation_fairness(data, variant, outcome, person = NULL)compare_presentation_fairness(data, variant, outcome, person = NULL)
data |
Data containing presentation version and an outcome. |
variant |
Presentation-version column. |
outcome |
Numeric outcome. |
person |
Optional participant column for descriptive aggregation. |
An object of class "eye_presentation_fairness_comparison", stored as a named list, with components "model", "summary", "status", "caveat". It contains outcomes across presentation variants and associated metadata or diagnostics needed to interpret the result.
Compare process external-validity models
compare_process_criterion_models(x)compare_process_criterion_models(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing process external-validity models. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compare explicit process-model specifications
compare_process_models( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )compare_process_models( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )
data |
Data. |
methods |
Named methods/specifications. |
analysis_fun |
Function '(data, method, specification)'. |
extract_fun |
Result extractor. |
An object of class "eye_process_sensitivity", stored as a named list, with components "grid", "results", "failures", "warnings", "grid_hash", "created_at", "status", "caveat". It contains explicit process-model specifications and associated metadata or diagnostics needed to interpret the result.
Compare candidate process-profile solutions
compare_process_profile_solutions(data, variables, k_values = 2:6, seed = 777)compare_process_profile_solutions(data, variables, k_values = 2:6, seed = 777)
data |
Person-level process data. |
variables |
Variables used for profiling. |
k_values |
Candidate numbers of profiles. |
seed |
Seed. |
A tabular R object containing candidate process-profile solutions; rows represent analysis units and columns contain the returned quantities.
Compare pupil deconvolution kernels
compare_pupil_kernels(data, tmax_values = c(512, 930), ...)compare_pupil_kernels(data, tmax_values = c(512, 930), ...)
data |
Same input used for fitting. |
tmax_values |
Candidate peak times. |
... |
Passed to 'fit_pupil_event_deconvolution()'. |
A tabular R object containing pupil deconvolution kernels; rows represent analysis units and columns contain the returned quantities.
Compare explicit pupil-preprocessing methods
compare_pupil_preprocessing( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )compare_pupil_preprocessing( data, methods, analysis_fun, extract_fun = .ep09_default_sensitivity_extract )
data |
Data. |
methods |
Named methods/specifications. |
analysis_fun |
Function '(data, method, specification)'. |
extract_fun |
Result extractor. |
An object of class "eye_process_sensitivity", stored as a named list, with components "grid", "results", "failures", "warnings", "grid_hash", "created_at", "status", "caveat". It contains explicit pupil-preprocessing methods and associated metadata or diagnostics needed to interpret the result.
Compare raw and confound-adjusted pupil values
compare_raw_adjusted_pupil(x)compare_raw_adjusted_pupil(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing raw and confound-adjusted pupil values. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compare two reproducibility fingerprints
compare_reproducibility_fingerprints(old, new)compare_reproducibility_fingerprints(old, new)
old, new
|
Fingerprints. |
A named list with components "detail", "identical", "old_hash", "new_hash", containing two reproducibility fingerprints and associated metadata or diagnostics.
Compare multiple signal filters
compare_signal_filters( signal, widths = c(5L, 9L, 15L), methods = c("runmed", "robfilter") )compare_signal_filters( signal, widths = c(5L, 9L, 15L), methods = c("runmed", "robfilter") )
signal |
Numeric signal. |
widths |
Widths to compare. |
methods |
Methods to compare. |
A tabular R object containing multiple signal filters; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_strategy_heterogeneity(object)compare_strategy_heterogeneity(object)
object |
Strategy fit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare validation engines on common recovery output
compare_validation_engines(results)compare_validation_engines(results)
results |
Validation or model results. |
An object of class "eye_validation_engine_comparison", "data.frame", stored as a data frame, containing validation engines on common recovery output and associated metadata needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
compare_vendor_semantics(x, vendors = NULL)compare_vendor_semantics(x, vendors = NULL)
x |
Corpus path or semantics data frame. |
vendors |
Optional vendor subset. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare base and visual-context IRT models
compare_visual_context_irt(x)compare_visual_context_irt(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing base and visual-context IRT models. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Build a detailed compatibility evidence matrix
compatibility_evidence_matrix(compatibility, evidence = NULL)compatibility_evidence_matrix(compatibility, evidence = NULL)
compatibility |
Existing output from 'build_compatibility_matrix()' or a compatible data frame. |
evidence |
Case-level evidence with columns 'ecosystem', 'device', 'evidence_level', and optionally 'semantic_roundtrip_pass'. |
An 'eye_compatibility_evidence_matrix' object.
Extract visual-context factor effects/loadings
context_factor_effects(x, IRTpars = FALSE)context_factor_effects(x, IRTpars = FALSE)
x |
Object to process, inspect, compare, or plot. |
IRTpars |
Passed to the underlying IRT coefficient extractor to request IRT parameterization when supported. |
An R object containing visual-context factor effects/loadings. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Detects lossless preservation and stable affine coordinate transformations.
coordinate_fidelity_audit( source, roundtrip, source_x = "x", source_y = "y", roundtrip_x = source_x, roundtrip_y = source_y, key = NULL, tolerance = 1e-06, correlation_floor = 0.999 )coordinate_fidelity_audit( source, roundtrip, source_x = "x", source_y = "y", roundtrip_x = source_x, roundtrip_y = source_y, key = NULL, tolerance = 1e-06, correlation_floor = 0.999 )
source |
Original/source representation. |
roundtrip |
Round-tripped or comparison representation. |
source_x |
Source horizontal coordinate column. |
source_y |
Source vertical coordinate column. |
roundtrip_x |
Round-tripped horizontal coordinate column. |
roundtrip_y |
Round-tripped vertical coordinate column. |
key |
Column or columns used to align records. |
tolerance |
Numerical tolerance used by the comparison. |
correlation_floor |
Minimum correlation treated as compatible. |
An object of class "eye_coordinate_fidelity", stored as a named list, with components "status", "x", "y", "matched_n", "tolerance", "correlation_floor". It contains coordinate semantic-fidelity audit and associated metadata or diagnostics needed to interpret the result.
Interval coverage calibration curve
coverage_calibration_curve(truth, lower, upper, nominal = NULL)coverage_calibration_curve(truth, lower, upper, nominal = NULL)
truth |
True values. |
lower |
Matrix/data.frame of lower limits or numeric vector. |
upper |
Matrix/data.frame of upper limits or numeric vector. |
nominal |
Nominal coverage labels, one per interval column. |
A data frame containing interval coverage calibration curve. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Create a public, de-identified benchmark bundle
create_public_benchmark( x, path, max_participants = 50L, include_samples = FALSE, overwrite = FALSE )create_public_benchmark( x, path, max_participants = 50L, include_samples = FALSE, overwrite = FALSE )
x |
An 'eye_dataset'. |
path |
Output directory. |
max_participants |
Optional participant cap. |
include_samples |
Whether to include sample-level tables. |
overwrite |
Whether to replace the output directory. |
Output directory.
Fits a simple affine linking map on anchor observations and reports residual bias/RMSE by device. It complements, rather than replaces, IRT anchor linking.
cross_device_process_equating_audit( data, value, reference_value, device, anchor = NULL )cross_device_process_equating_audit( data, value, reference_value, device, anchor = NULL )
data |
Input data frame or compatible tabular object. |
value |
Process-value column or values. |
reference_value |
Value supplied to 'reference_value'; see Details for its model-specific role. |
device |
Device identifier or device facet. |
anchor |
Anchor or reference group used for linking. |
An object of class "eye_cross_device_equating_audit", "data.frame", stored as a data frame, containing cross-device process-scale equating audit and associated metadata needed to interpret the result.
Compare adapter output across software/format versions
cross_version_adapter_regression( input, baseline_adapter, candidate_adapter, baseline_version = "baseline", candidate_version = "candidate", extract_samples = .ep07_adapter_extract_samples, audit_args = list() )cross_version_adapter_regression( input, baseline_adapter, candidate_adapter, baseline_version = "baseline", candidate_version = "candidate", extract_samples = .ep07_adapter_extract_samples, audit_args = list() )
input |
Shared raw fixture/input. |
baseline_adapter, candidate_adapter
|
Functions that parse 'input'. |
baseline_version, candidate_version
|
Version labels. |
extract_samples |
Function extracting comparable sample tables. |
audit_args |
Arguments passed to 'field_fidelity_report()'. |
An object of class "eye_adapter_regression_audit", stored as a named list, with components "status", "baseline_version", "candidate_version", "fidelity", "baseline", "candidate". It contains adapter output across software/format versions and associated metadata or diagnostics needed to interpret the result.
Cross-classified grouped cross-validation
crossed_grouped_cv( data, formula, family = stats::binomial(), groups = c("participant_id", "item_id"), v = 5L, metric = c("log_loss", "brier", "accuracy"), seed = 1L )crossed_grouped_cv( data, formula, family = stats::binomial(), groups = c("participant_id", "item_id"), v = 5L, metric = c("log_loss", "brier", "accuracy"), seed = 1L )
data |
Data frame. |
formula |
Model formula. |
family |
GLM family. |
groups |
Crossed grouping columns. |
v |
Number of folds. |
metric |
Metric: log loss, Brier score, or accuracy. |
seed |
Random seed. |
An 'eye_crossed_grouped_cv' object.
Holds out levels from every declared grouping dimension simultaneously. The assessment set is the intersection of held-out levels; the analysis set excludes every held-out level. Rows combining held-out and retained levels form a buffer and are deliberately used in neither set.
crossed_grouped_folds( data, groups = c("participant_id", "item_id"), v = 5L, seed = 1L )crossed_grouped_folds( data, groups = c("participant_id", "item_id"), v = 5L, seed = 1L )
data |
Data frame. |
groups |
Two or more crossed grouping columns. |
v |
Number of folds. |
seed |
Random seed. |
An 'eye_crossed_grouped_folds' object.
Compact reporting table for eye-tracking data quality
data_quality_reporting_table(x)data_quality_reporting_table(x)
x |
Data-quality profile. |
An R object containing compact reporting table for eye-tracking data quality. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Alias for manifest comparison emphasizing changed decision paths
decision_manifest_diff(old, new)decision_manifest_diff(old, new)
old |
Earlier manifest. |
new |
Later manifest. |
An R object containing alias for manifest comparison emphasizing changed decision paths. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Stable hash of decision content
decision_manifest_hash(x)decision_manifest_hash(x)
x |
Manifest or decision object. |
An R object containing stable hash of decision content. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Flatten a decision manifest to a table
decision_manifest_table(x)decision_manifest_table(x)
x |
Manifest. |
A logical value or vector indicating flatten a decision manifest to a table.
Coverage of a declared decision space by evaluated specifications
decision_space_coverage(grid, evaluated)decision_space_coverage(grid, evaluated)
grid |
Sensitivity grid. |
evaluated |
Sensitivity result or vector of evaluated specification IDs. |
A data frame containing coverage of a declared decision space by evaluated specifications. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Overall decision-stability summary
decision_stability( x, effect = "effect", p_value = NULL, alpha = 0.05, threshold = 0 )decision_stability( x, effect = "effect", p_value = NULL, alpha = 0.05, threshold = 0 )
x |
Sensitivity result. |
effect |
Effect column. |
p_value |
Optional p-value column. |
alpha |
Significance threshold. |
threshold |
Substantive threshold. |
An object of class "eye_decision_stability", stored as a named list, with components "summary", "stable_sign", "stable_threshold", "stable_significance", "thresholds", "caveat". It contains overall decision-stability summary and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
decode_dynamic_states(object, method = c("mode", "probability", "draw"))decode_dynamic_states(object, method = c("mode", "probability", "draw"))
object |
Dynamic IRTree fit. |
method |
Marginal probabilities, posterior mode, or draw. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compositional analysis of AOI attention. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
derive_aoi_composition(x, aois, denominator = c("total_aoi_dwell", "trial_duration"), zero_method = c("multiplicative", "bayesian"), id_cols = NULL, aoi_col = "aoi", value_col = "dwell_ms", trial_duration_col = NULL) transform_aoi_composition(x, method = c("ilr", "clr", "alr"), reference = NULL) fit_aoi_compositional_model(composition, formula, random = NULL, data = NULL, method = "ilr") compare_aoi_compositions(x, group, method = c("permanova", "compositional_manova"), permutations = 499, seed = 20260807) aoi_balance_coordinates(x, balances) plot_aoi_ternary(x, ...) plot_aoi_balance_biplot(x, ...) plot_aoi_variation_matrix(x, ...) plot_compositional_group_difference(x, ...) plot_aoi_composition_trajectory(x, ...)derive_aoi_composition(x, aois, denominator = c("total_aoi_dwell", "trial_duration"), zero_method = c("multiplicative", "bayesian"), id_cols = NULL, aoi_col = "aoi", value_col = "dwell_ms", trial_duration_col = NULL) transform_aoi_composition(x, method = c("ilr", "clr", "alr"), reference = NULL) fit_aoi_compositional_model(composition, formula, random = NULL, data = NULL, method = "ilr") compare_aoi_compositions(x, group, method = c("permanova", "compositional_manova"), permutations = 499, seed = 20260807) aoi_balance_coordinates(x, balances) plot_aoi_ternary(x, ...) plot_aoi_balance_biplot(x, ...) plot_aoi_variation_matrix(x, ...) plot_compositional_group_difference(x, ...) plot_aoi_composition_trajectory(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
aois |
Argument controlling 'aois'; see the function usage and returned audit metadata. |
denominator |
Argument controlling 'denominator'; see the function usage and returned audit metadata. |
zero_method |
Argument controlling 'zero_method'; see the function usage and returned audit metadata. |
id_cols |
Argument controlling 'id_cols'; see the function usage and returned audit metadata. |
aoi_col |
Argument controlling 'aoi_col'; see the function usage and returned audit metadata. |
value_col |
Argument controlling 'value_col'; see the function usage and returned audit metadata. |
trial_duration_col |
Argument controlling 'trial_duration_col'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
reference |
Argument controlling 'reference'; see the function usage and returned audit metadata. |
composition |
Argument controlling 'composition'; see the function usage and returned audit metadata. |
formula |
Argument controlling 'formula'; see the function usage and returned audit metadata. |
random |
Argument controlling 'random'; see the function usage and returned audit metadata. |
data |
Argument controlling 'data'; see the function usage and returned audit metadata. |
group |
Argument controlling 'group'; see the function usage and returned audit metadata. |
permutations |
Argument controlling 'permutations'; see the function usage and returned audit metadata. |
seed |
Argument controlling 'seed'; see the function usage and returned audit metadata. |
balances |
Argument controlling 'balances'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Derive the complete Gazepoint workflow feature set
derive_gazepoint_workflow_features(x, reset_workflow_features = TRUE)derive_gazepoint_workflow_features(x, reset_workflow_features = TRUE)
x |
An 'eye_dataset' with reconstructed media trials. |
reset_workflow_features |
Remove features previously generated by this workflow while preserving native Gazepoint Data Summary features. |
The updated 'eye_dataset'.
Calibration drift and offline recalibration. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
detect_calibration_drift(x, references = NULL, window = "30 sec", method = c("targets", "known_aois", "fixation_density"), x_col = NULL, y_col = NULL, time_col = NULL) fit_offline_recalibration(x, method = c("translation", "affine", "polynomial"), robust = TRUE, x_col = NULL, y_col = NULL, reference_x_col = NULL, reference_y_col = NULL) apply_offline_recalibration(x, model, x_col = NULL, y_col = NULL, suffix = "_recalibrated") audit_recalibration(before, after, minimum_improvement = NULL) plot_calibration_vector_field(x, ...) plot_calibration_error_ellipses(x, ...) plot_drift_over_time(x, ...) plot_recalibration_before_after(x, ...) plot_screen_coverage(x, ...)detect_calibration_drift(x, references = NULL, window = "30 sec", method = c("targets", "known_aois", "fixation_density"), x_col = NULL, y_col = NULL, time_col = NULL) fit_offline_recalibration(x, method = c("translation", "affine", "polynomial"), robust = TRUE, x_col = NULL, y_col = NULL, reference_x_col = NULL, reference_y_col = NULL) apply_offline_recalibration(x, model, x_col = NULL, y_col = NULL, suffix = "_recalibrated") audit_recalibration(before, after, minimum_improvement = NULL) plot_calibration_vector_field(x, ...) plot_calibration_error_ellipses(x, ...) plot_drift_over_time(x, ...) plot_recalibration_before_after(x, ...) plot_screen_coverage(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
references |
Argument controlling 'references'; see the function usage and returned audit metadata. |
window |
Argument controlling 'window'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
x_col |
Argument controlling 'x_col'; see the function usage and returned audit metadata. |
y_col |
Argument controlling 'y_col'; see the function usage and returned audit metadata. |
time_col |
Argument controlling 'time_col'; see the function usage and returned audit metadata. |
robust |
Argument controlling 'robust'; see the function usage and returned audit metadata. |
reference_x_col |
Argument controlling 'reference_x_col'; see the function usage and returned audit metadata. |
reference_y_col |
Argument controlling 'reference_y_col'; see the function usage and returned audit metadata. |
model |
Argument controlling 'model'; see the function usage and returned audit metadata. |
suffix |
Argument controlling 'suffix'; see the function usage and returned audit metadata. |
before |
Argument controlling 'before'; see the function usage and returned audit metadata. |
after |
Argument controlling 'after'; see the function usage and returned audit metadata. |
minimum_improvement |
Argument controlling 'minimum_improvement'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
detect_corrupt_partitions(storage)detect_corrupt_partitions(storage)
storage |
Storage object or path. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
This implementation is a transparent package reference inspired by the 2026 SIC-CPA literature. It combines Bernoulli response likelihood with normal log-RT and standardized gaze likelihoods. It is not a line-for-line reproduction of the article's estimator.
detect_irt_changepoints( data, person = "participant_id", order = "item_order", response = "response", rt = "rt", gaze = NULL, min_segment = 5L, min_delta_sic = 2, max_changes = 2L )detect_irt_changepoints( data, person = "participant_id", order = "item_order", response = "response", rt = "rt", gaze = NULL, min_segment = 5L, min_delta_sic = 2, max_changes = 2L )
data |
Input data frame or compatible tabular object. |
person |
Person or participant identifier column. |
order |
Within-sequence ordering variable. |
response |
Response variable or response-column name. |
rt |
Response-time variable or column name. |
gaze |
Gaze/process variable or column name. |
min_segment |
Minimum segment length. |
min_delta_sic |
Minimum information-criterion improvement. |
max_changes |
Maximum number of change points. |
An object of class "eye_irt_changepoints", stored as a named list, with components "results", "channels", "method", "min_segment", "min_delta_sic", "max_changes". It contains iRT/process change points using an SIC-inspired multichannel score and associated metadata or diagnostics needed to interpret the result.
Public roadmap alias for 'detect_irt_changepoints()'.
detect_process_changepoint(...)detect_process_changepoint(...)
... |
Additional arguments passed to the selected model, engine, or method. |
An R object containing a response-process change point. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Cognitive-episode change-point detection. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
detect_process_changepoints(x, channels = c("gaze_velocity", "aoi", "pupil", "eda"), time_col = NULL, window = 10, threshold_quantile = 0.9, min_segment = 5) segment_process_episodes(x, ...) label_process_episodes(x, rules = NULL, model = NULL) compare_episode_structure(x, group) plot_process_episodes(x, ...) plot_changepoint_ribbons(x, ...) plot_episode_waterfall(x, ...) plot_episode_transition_graph(x, ...) plot_episode_duration_distribution(x, ...)detect_process_changepoints(x, channels = c("gaze_velocity", "aoi", "pupil", "eda"), time_col = NULL, window = 10, threshold_quantile = 0.9, min_segment = 5) segment_process_episodes(x, ...) label_process_episodes(x, rules = NULL, model = NULL) compare_episode_structure(x, group) plot_process_episodes(x, ...) plot_changepoint_ribbons(x, ...) plot_episode_waterfall(x, ...) plot_episode_transition_graph(x, ...) plot_episode_duration_distribution(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
channels |
Argument controlling 'channels'; see the function usage and returned audit metadata. |
time_col |
Argument controlling 'time_col'; see the function usage and returned audit metadata. |
window |
Argument controlling 'window'; see the function usage and returned audit metadata. |
threshold_quantile |
Argument controlling 'threshold_quantile'; see the function usage and returned audit metadata. |
min_segment |
Argument controlling 'min_segment'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
rules |
Argument controlling 'rules'; see the function usage and returned audit metadata. |
model |
Argument controlling 'model'; see the function usage and returned audit metadata. |
group |
Argument controlling 'group'; see the function usage and returned audit metadata. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Extract device facet effects
device_facet_effects(object, channel = c("response", "process"))device_facet_effects(object, channel = c("response", "process"))
object |
A fitted eyeprocess model or audit object. |
channel |
Measurement channel to inspect. |
An object of class "eye_process_facet_effects", stored as a named list, with components "facet", "column", "channel", "random_effects", "variance_component". It contains device facet effects and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
diffusion_identification_study(conditions = list(n_person = c(50L, 150L), n_item = c(10L, 30L), gaze_effect = c(0, 0.35), contaminant_fraction = c(0, 0.05)), replications = 20L, base_seed = 20260805L, spec = gaze_diffusion_spec(drift_features = "gaze_balance", engine = "stan"))diffusion_identification_study(conditions = list(n_person = c(50L, 150L), n_item = c(10L, 30L), gaze_effect = c(0, 0.35), contaminant_fraction = c(0, 0.05)), replications = 20L, base_seed = 20260805L, spec = gaze_diffusion_spec(drift_features = "gaze_balance", engine = "stan"))
conditions |
Named list or data frame of design conditions. |
replications |
Replications per condition. |
base_seed |
Base seed. |
spec |
Confirmatory diffusion specification used for each fit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
diffusion_parameter_diagnostics(object, correlation_threshold = 0.85)diffusion_parameter_diagnostics(object, correlation_threshold = 0.85)
object |
Diffusion fit. |
correlation_threshold |
Correlation threshold. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
diffusion_posterior_predictive(object, draws = 200L, method = c("rtdists", "stan_proxy"), seed = 1L)diffusion_posterior_predictive(object, draws = 200L, method = c("rtdists", "stan_proxy"), seed = 1L)
object |
Diffusion fit. |
draws |
Number of generated-quantity draws to retain. |
method |
Value for 'method'. See the function description and relevant article for constraints. |
seed |
Value for 'seed'. See the function description and relevant article for constraints. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Build a distractor process map
distractor_process_map(object)distractor_process_map(object)
object |
A fitted eyeprocess model or audit object. |
A data frame containing a distractor process map. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Drift audit stratified by device
drift_by_device(data, device = "device_id", ...)drift_by_device(data, device = "device_id", ...)
data |
Data frame containing the required process variables. |
device |
Name of the column identifying device. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing drift audit stratified by device. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Drift audit stratified by study site
drift_by_site(data, site = "site_id", ...)drift_by_site(data, site = "site_id", ...)
data |
Data frame containing the required process variables. |
site |
Name of the column identifying site. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing drift audit stratified by study site. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Drift audit stratified by stimulus version
drift_by_stimulus_version(data, stimulus_version = "stimulus_version", ...)drift_by_stimulus_version(data, stimulus_version = "stimulus_version", ...)
data |
Data frame containing the required process variables. |
stimulus_version |
Name of the column identifying stimulus version. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing drift audit stratified by stimulus version. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Drift audit stratified by vendor
drift_by_vendor(data, vendor = "vendor", ...)drift_by_vendor(data, vendor = "vendor", ...)
data |
Data frame containing the required process variables. |
vendor |
Name of the column identifying vendor. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing drift audit stratified by vendor. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
dynamic_irtree_recovery(grid = expand.grid(state_misclassification = c(0, 0.05, 0.15), missing_state = c(0, 0.10), stringsAsFactors = FALSE), replications = 20L, spec = dynamic_irtree_spec(engine = "multinomial"), base_seed = 1L)dynamic_irtree_recovery(grid = expand.grid(state_misclassification = c(0, 0.05, 0.15), missing_state = c(0, 0.10), stringsAsFactors = FALSE), replications = 20L, spec = dynamic_irtree_spec(engine = "multinomial"), base_seed = 1L)
grid |
Scenario grid. |
replications |
Replications per scenario. |
spec |
Dynamic IRTree specification. |
base_seed |
Base seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
dynamic_irtree_spec(source = c("samples", "visits", "fixations"), collapse_consecutive = TRUE, engine = c("baseline", "multinomial", "stan"), hidden_states = 0L, include_response = TRUE, include_person = FALSE, include_item = TRUE, condition_columns = character(), transition_predictors = character(), interactions = character(), include_time_gap = TRUE, person_effect = c("none", "fixed", "random"), item_effect = c("none", "fixed", "random"), structural_zeros = NULL, allowed_transitions = NULL, hidden_structural_zeros = NULL, hidden_allowed_transitions = NULL, missing_state = c("drop", "unknown", "marginalize"), uncertain_state_probability = NULL, misclassification_matrix = NULL, ridge = 1e-4, standardize = TRUE, reference_state = NULL, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L)dynamic_irtree_spec(source = c("samples", "visits", "fixations"), collapse_consecutive = TRUE, engine = c("baseline", "multinomial", "stan"), hidden_states = 0L, include_response = TRUE, include_person = FALSE, include_item = TRUE, condition_columns = character(), transition_predictors = character(), interactions = character(), include_time_gap = TRUE, person_effect = c("none", "fixed", "random"), item_effect = c("none", "fixed", "random"), structural_zeros = NULL, allowed_transitions = NULL, hidden_structural_zeros = NULL, hidden_allowed_transitions = NULL, missing_state = c("drop", "unknown", "marginalize"), uncertain_state_probability = NULL, misclassification_matrix = NULL, ridge = 1e-4, standardize = TRUE, reference_state = NULL, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L)
source |
Sequence source for an 'eye_dataset'. |
collapse_consecutive |
Collapse consecutive identical states. |
engine |
Estimation engine: auditable binary baseline, penalized |
|
Number of latent states. Zero fits observed-state |
|
include_response |
Include item response as a predictor. |
include_person |
Include person effects. |
include_item |
Include item effects. |
condition_columns |
Condition-level predictors. |
transition_predictors |
Additional transition-level predictors. |
interactions |
Optional interaction terms supplied as formula strings. |
include_time_gap |
Include log time gap for irregular observations. |
person_effect |
Person effect type for Stan. |
item_effect |
Item effect type for Stan. |
structural_zeros |
Optional forbidden observed-state transition pairs. |
allowed_transitions |
Optional allowed observed-state transition pairs. |
|
Optional forbidden hidden-state pairs named 'state1', 'state2', and so forth. |
|
|
Optional explicitly allowed hidden-state pairs. |
|
missing_state |
Treatment of missing observed states. |
uncertain_state_probability |
Optional column containing probability of |
misclassification_matrix |
Optional observed-state misclassification matrix. |
ridge |
Penalization for the multinomial baseline. |
standardize |
Standardize numeric design columns. |
reference_state |
Optional reference destination state. |
chains |
CmdStan controls. |
parallel_chains |
CmdStan controls. |
iter_warmup |
CmdStan controls. |
iter_sampling |
CmdStan controls. |
adapt_delta |
CmdStan sampler controls. |
max_treedepth |
CmdStan sampler controls. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
dynamic_posterior_predictive_check(object, draws = 200L, seed = 1L)dynamic_posterior_predictive_check(object, draws = 200L, seed = 1L)
object |
Dynamic IRTree fit using the Stan engine. |
draws |
Maximum posterior predictive draws to summarize. |
seed |
Seed used when subsampling posterior draws. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
dynamic_transition_design(data, spec = dynamic_irtree_spec(), formula = NULL)dynamic_transition_design(data, spec = dynamic_irtree_spec(), formula = NULL)
data |
Prepared transition data. |
spec |
Dynamic IRTree specification. |
formula |
Optional explicit right-hand-side formula. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Estimate effective sampling frequency from timestamps
effective_sampling_frequency( data, time = "timestamp_ms", unit = c("ms", "s", "us"), by = NULL )effective_sampling_frequency( data, time = "timestamp_ms", unit = c("ms", "s", "us"), by = NULL )
data |
Sample data. |
time |
Timestamp column. |
unit |
Timestamp unit. |
by |
Optional grouping columns. |
A logical value or vector indicating effective sampling frequency from timestamps.
Encode multiple-response item response combinations
encode_response_combinations( data, person = "participant_id", item = "item_id", option = "option_id", selected = "selected", sort_options = TRUE, empty_code = "<none>" )encode_response_combinations( data, person = "participant_id", item = "item_id", option = "option_id", selected = "selected", sort_options = TRUE, empty_code = "<none>" )
data |
Long person-item-option table. |
person, item, option, selected
|
Column names. |
sort_options |
Sort selected option labels before combining. |
empty_code |
Code for no selected options. |
A tabular R object containing encode multiple-response item response combinations; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
engine_adapter_status(engine)engine_adapter_status(engine)
engine |
Engine name. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Implements mean-sigma, mean-mean, Stocking-Lord, and Haebara linking for dichotomous 2PL-style item parameters. New-form parameters are transformed onto the reference scale using theta_ref = A * theta_new + B.
equate_irt_scales( reference, new, method = c("stocking-lord", "haebara", "mean-sigma", "mean-mean"), theta_grid = seq(-4, 4, length.out = 81) )equate_irt_scales( reference, new, method = c("stocking-lord", "haebara", "mean-sigma", "mean-mean"), theta_grid = seq(-4, 4, length.out = 81) )
reference |
Reference-scale parameters or data. |
new |
New-scale parameters or data. |
method |
Method used for estimation, linking, or comparison. |
theta_grid |
Grid of latent-trait values used for evaluation. |
An object of class "eye_irt_equating", stored as a named list, with components "A", "B", "method", "transformed", "reference", "new", "equation". It contains equate IRT scales using anchor item parameters and associated metadata or diagnostics needed to interpret the result.
Estimate empirical calibration/validation error
estimate_calibration_error( data, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y", by = NULL )estimate_calibration_error( data, gaze_x = "gaze_x", gaze_y = "gaze_y", target_x = "target_x", target_y = "target_y", by = NULL )
data |
Validation-target data. |
gaze_x, gaze_y
|
Recorded gaze-coordinate columns. |
target_x, target_y
|
Known target-coordinate columns. |
by |
Optional grouping columns such as participant/session. |
A tabular R object containing empirical calibration/validation error; rows represent analysis units and columns contain the returned quantities.
Estimate visual exposure probability
estimate_visual_exposure_probability( data, exposed = "reached", predictors, family = stats::binomial() )estimate_visual_exposure_probability( data, exposed = "reached", predictors, family = stats::binomial() )
data |
Input data frame or compatible tabular object. |
exposed |
Value supplied to 'exposed'; see Details for its model-specific role. |
predictors |
Predictor variables used by the model. |
family |
Statistical family used by the channel or model. |
An object of class "eye_visual_exposure_model", stored as a named list, with components "model", "fitted_probability", "exposed", "predictors". It contains visual exposure probability and associated metadata or diagnostics needed to interpret the result.
Evaluate a validation acceptance rule
evaluate_validation_acceptance(value, rule)evaluate_validation_acceptance(value, rule)
value |
Observed metric value to evaluate. |
rule |
Validation acceptance rule to apply. |
A logical value or vector indicating a validation acceptance rule.
Build explicit event-detector specifications, run detector branches independently, compare event catalogues, propagate detector choice through AOI features and an identical statistical model, and report scientific inference stability.
define_event_detector_spec( detector_id, algorithm, velocity_threshold = NULL, dispersion_threshold = NULL, minimum_duration_ms = NULL, maximum_gap_ms = NULL, merge_rule = "none", sampling_rate = NULL, smoothing = NULL, filter = NULL, coordinate_unit = "degrees", implementation = NULL, implementation_version = NULL, parameters = list(), callback = NULL ) validate_event_detector_spec(spec) create_detector_multiverse( specs = NULL, base_spec = NULL, parameter_grid = NULL, id_template = "%s_%03d", label = "event_detector_multiverse" ) detect_events_with_spec(x, spec) import_external_detector_events(events, spec, dataset = NULL) run_detector_multiverse(x, multiverse, continue_on_error = TRUE) match_detected_events( reference, candidate, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) compare_event_catalogues( reference, candidate, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) estimate_detector_agreement( x, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) summarise_detector_events(x) summarise_detector_disagreement(x, event_type = "fixation") propagate_detector_to_aoi(x, overlap = "error", continue_on_error = TRUE) propagate_detector_to_features(x, continue_on_error = TRUE) run_detector_inference_multiverse( x, model_spec, model_callback = NULL, minimum_valid_fraction = NULL ) assess_detector_inference_stability( x, term, substantive_threshold = NULL, direction = c("above", "below", "absolute") ) summarise_detector_robustness( x, inference = NULL, term = NULL, substantive_threshold = NULL, direction = "above" ) plot_detector_event_timeline(x, trial_id = NULL, event_type = "fixation", ...) plot_detector_agreement(x, metric = "mean_event_overlap", ...) plot_detector_feature_distributions( x, feature = "dwell_time_ms", aoi_id = NULL, ... ) plot_detector_coefficient_stability(x, term, ...) plot_detector_multiverse( x, inference = NULL, term = NULL, feature = "dwell_time_ms", aoi_id = NULL ) report_detector_multiverse( x, inference = NULL, term = NULL, substantive_threshold = NULL, path = NULL ) simulate_detector_multiverse_data( n_participants = 12L, sampling_rate = 60, trial_duration_s = 2, seed = 20260918L )define_event_detector_spec( detector_id, algorithm, velocity_threshold = NULL, dispersion_threshold = NULL, minimum_duration_ms = NULL, maximum_gap_ms = NULL, merge_rule = "none", sampling_rate = NULL, smoothing = NULL, filter = NULL, coordinate_unit = "degrees", implementation = NULL, implementation_version = NULL, parameters = list(), callback = NULL ) validate_event_detector_spec(spec) create_detector_multiverse( specs = NULL, base_spec = NULL, parameter_grid = NULL, id_template = "%s_%03d", label = "event_detector_multiverse" ) detect_events_with_spec(x, spec) import_external_detector_events(events, spec, dataset = NULL) run_detector_multiverse(x, multiverse, continue_on_error = TRUE) match_detected_events( reference, candidate, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) compare_event_catalogues( reference, candidate, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) estimate_detector_agreement( x, event_type = "fixation", onset_tolerance_ms = 75, minimum_overlap = 0.10 ) summarise_detector_events(x) summarise_detector_disagreement(x, event_type = "fixation") propagate_detector_to_aoi(x, overlap = "error", continue_on_error = TRUE) propagate_detector_to_features(x, continue_on_error = TRUE) run_detector_inference_multiverse( x, model_spec, model_callback = NULL, minimum_valid_fraction = NULL ) assess_detector_inference_stability( x, term, substantive_threshold = NULL, direction = c("above", "below", "absolute") ) summarise_detector_robustness( x, inference = NULL, term = NULL, substantive_threshold = NULL, direction = "above" ) plot_detector_event_timeline(x, trial_id = NULL, event_type = "fixation", ...) plot_detector_agreement(x, metric = "mean_event_overlap", ...) plot_detector_feature_distributions( x, feature = "dwell_time_ms", aoi_id = NULL, ... ) plot_detector_coefficient_stability(x, term, ...) plot_detector_multiverse( x, inference = NULL, term = NULL, feature = "dwell_time_ms", aoi_id = NULL ) report_detector_multiverse( x, inference = NULL, term = NULL, substantive_threshold = NULL, path = NULL ) simulate_detector_multiverse_data( n_participants = 12L, sampling_rate = 60, trial_duration_s = 2, seed = 20260918L )
detector_id |
Unique detector-branch identifier. |
algorithm |
One of |
velocity_threshold |
Explicit velocity threshold in the declared coordinate unit per second. |
dispersion_threshold |
Explicit dispersion threshold in the declared coordinate unit. |
minimum_duration_ms |
Minimum event duration in milliseconds. |
maximum_gap_ms |
Maximum gap used by algorithms that support gap handling. |
merge_rule |
Recorded merge rule. No undeclared merging is applied. |
sampling_rate |
Declared sample rate in Hz. |
smoothing, filter
|
Recorded preprocessing choices. |
coordinate_unit |
|
implementation, implementation_version
|
Detector implementation identity and version. |
parameters |
Named algorithm-specific parameter list. |
callback |
Explicit external detector or model callback. |
spec |
An event-detector specification. |
specs |
A list of explicit detector specifications. |
base_spec |
Base specification for an explicit parameter grid. |
parameter_grid |
Named list of parameter values. Values are study-specific alternatives, not universal defaults. |
id_template |
|
label |
Multiverse label. |
x |
An |
multiverse |
An event-detector multiverse created by
|
events, reference, candidate
|
Canonical or canonicalizable event data frames. |
dataset |
Optional source |
continue_on_error |
Whether to retain a failed branch and continue other specifications. |
event_type |
Event type to compare, usually |
onset_tolerance_ms |
Allowed onset difference for near-boundary matches. |
minimum_overlap |
Minimum temporal intersection-over-union for event matching. |
overlap |
AOI-overlap rule. The default |
model_spec |
Named list declaring the model engine, formula, outcome, and optional AOI/model settings. |
model_callback |
Specialist model callback returning the documented tidy coefficient contract. |
minimum_valid_fraction |
Explicit minimum gaze-validity fraction before modelling. |
term |
Model coefficient term to assess. |
substantive_threshold |
Optional analyst-declared effect threshold. |
direction |
How the substantive threshold is interpreted. |
inference |
Optional detector inference result. |
trial_id |
Optional trial filter. |
metric |
Agreement metric to plot. |
feature |
Propagated feature to plot. |
aoi_id |
Optional AOI filter. |
path |
Optional Markdown report path. |
n_participants |
Synthetic-example participant count. |
trial_duration_s |
Synthetic trial duration in seconds. |
seed |
Synthetic-data random seed. |
... |
Additional base-graphics arguments. |
The multiverse is intended to answer whether event detection materially changes the quantities and scientific inference used by a study. It is not a mechanism for selecting the branch with a preferred p-value.
Zero-event trials remain represented during feature propagation. Trials with no
valid gaze remain missing rather than being converted to zero. AOI ambiguity is
an error unless an analyst explicitly selects another overlap rule. Model engines
are explicit; formula engines use na.fail; and non-converged model branches
are retained diagnostically but excluded from coefficient-stability summaries.
algorithm = "remodnav" calls an installed REMoDNaV executable and never
substitutes an internal approximation. Pixel coordinates require an explicit
parameters$px2deg. The built-in adaptive-velocity implementation is
reported as a robust-MAD reference detector and must not be described as a named
external algorithm.
Specification functions return detector specification or multiverse objects.
Execution/propagation functions return governed multiverse result objects.
run_detector_inference_multiverse() returns an inference result containing
coefficients, failures, warnings, model metadata, a feature fingerprint, and
input_audit. The audit has one row per planned detector and records
input_rows, aoi_selected_rows, quality_excluded_rows,
outcome_missing_rows, model_rows_used, and branch status.
Comparison and stability functions return data frames or named summary lists.
For assess_detector_inference_stability(), specifications is the
number of planned detector specifications, term_available_specifications
is the number that returned the requested coefficient, and
model_failure_specifications is the number with recorded model-stage
failures. The convergence-rate denominator is the full declared multiverse, so
failed, missing-term, and non-converged branches cannot disappear from robustness
accounting.
Plotting functions draw base-R diagnostics and invisibly return plotted data.
report_detector_multiverse() returns Markdown text.
simulate_detector_multiverse_data() returns a synthetic eye_dataset.
detect_fixations_ivt, detect_fixations_idt,
assign_aois, preprocessing_multiverse
d <- simulate_detector_multiverse_data(n_participants = 4, seed = 1) s1 <- define_event_detector_spec( "ivt25", "ivt", velocity_threshold = 25, minimum_duration_ms = 60, maximum_gap_ms = 75, sampling_rate = 60 ) s2 <- define_event_detector_spec( "ivt35", "ivt", velocity_threshold = 35, minimum_duration_ms = 60, maximum_gap_ms = 75, sampling_rate = 60 ) mv <- create_detector_multiverse(list(s1, s2)) out <- run_detector_multiverse(d, mv) out <- propagate_detector_to_aoi(out) out <- propagate_detector_to_features(out) summarise_detector_events(out)d <- simulate_detector_multiverse_data(n_participants = 4, seed = 1) s1 <- define_event_detector_spec( "ivt25", "ivt", velocity_threshold = 25, minimum_duration_ms = 60, maximum_gap_ms = 75, sampling_rate = 60 ) s2 <- define_event_detector_spec( "ivt35", "ivt", velocity_threshold = 35, minimum_duration_ms = 60, maximum_gap_ms = 75, sampling_rate = 60 ) mv <- create_detector_multiverse(list(s1, s2)) out <- run_detector_multiverse(d, mv) out <- propagate_detector_to_aoi(out) out <- propagate_detector_to_features(out) summarise_detector_events(out)
Evaluates whether independent channel offsets corroborate a nominal event. This is annotation/event plausibility QC, not clock synchronization, and it never modifies timestamps.
event_marker_qc(offsets, tolerance, min_corroborating = 2L)event_marker_qc(offsets, tolerance, min_corroborating = 2L)
offsets |
Numeric corroborating-channel offsets in seconds. |
tolerance |
Allowed absolute offset in seconds. |
min_corroborating |
Minimum corroborating channels for 'confirmed'. |
A list containing status, consensus offset, uncertainty, and counts.
Audit event survival across an interchange round trip
event_roundtrip_audit(source_events, roundtrip_events, hed_column = NULL, ...)event_roundtrip_audit(source_events, roundtrip_events, hed_column = NULL, ...)
source_events, roundtrip_events
|
Event tables. |
hed_column |
Optional HED annotation column to compare structurally. |
... |
Arguments forwarded to 'event_semantics_audit()'. |
An object of class "eye_event_roundtrip_audit", stored as a named list, with components "status", "event_semantics", "hed". It contains event survival across an interchange round trip and associated metadata or diagnostics needed to interpret the result.
Audit event semantic preservation
event_semantics_audit( source_events, roundtrip_events, label = "event", time = "timestamp", key = NULL, tolerance = 1e-06 )event_semantics_audit( source_events, roundtrip_events, label = "event", time = "timestamp", key = NULL, tolerance = 1e-06 )
source_events, roundtrip_events
|
Event tables. |
label |
Event-label field. |
time |
Event-time field; set 'NULL' to compare labels only. |
key |
Optional event identity key. |
tolerance |
Timestamp tolerance. |
An object of class "eye_event_semantics", stored as a named list, with components "status", "source_n", "roundtrip_n", "matched_n", "exact_label_fraction", "max_time_error". It contains event semantic preservation and associated metadata or diagnostics needed to interpret the result.
Expand a stress evidence plan into one-factor-at-a-time scenarios
expand_eyeprocess_stress_evidence_plan(plan)expand_eyeprocess_stress_evidence_plan(plan)
plan |
Validation or stress-evidence plan object. |
A tabular R object containing expand a stress evidence plan into one-factor-at-a-time scenarios; rows represent analysis units and columns contain the returned quantities.
Expand a validation-evidence plan to a scenario table
expand_eyeprocess_validation_plan(x)expand_eyeprocess_validation_plan(x)
x |
Object to validate, summarize, verify, or otherwise process. |
An R object containing expand a validation-evidence plan to a scenario table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Expand a process-validation design into explicit conditions
expand_process_validation_design(x, max_conditions = 250000L)expand_process_validation_design(x, max_conditions = 250000L)
x |
Validation design. |
max_conditions |
Optional hard cap for accidental combinatorial explosion. |
A tabular R object containing expand a process-validation design into explicit conditions; rows represent analysis units and columns contain the returned quantities.
Expected process-aware item utility under a theta distribution
expected_process_information(info, theta_weights = NULL)expected_process_information(info, theta_weights = NULL)
info |
Value supplied to 'info'; see Details for its model-specific role. |
theta_weights |
Weights over the theta distribution. |
An R object containing expected process-aware item utility under a theta distribution. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Returns the closest person-item pairs in the fitted residual latent space. Closeness is descriptive residual structure, not a causal explanation.
explain_latent_interaction(object, person = NULL, item = NULL, top = 10L)explain_latent_interaction(object, person = NULL, item = NULL, top = 10L)
object |
Fitted 'eye_latent_space_irt' object. |
person |
Optional person row/index/name to restrict. |
item |
Optional item row/index/name to restrict. |
top |
Number of closest pairs to return. |
An R object containing explain local person-item latent-space interactions. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Writes BIDS-compatible '_physio.tsv.gz' and JSON pairs with 'PhysioType = "eyetrack"'. Continuous files contain no header; 'Columns' is stored in the sidecar. Each recorded eye is written separately.
export_eye_bids( x, path, task = "task", dataset_name = "eyeprocess eye-tracking dataset", overwrite = FALSE, screen_distance_m = NULL, screen_size_m = NULL )export_eye_bids( x, path, task = "task", dataset_name = "eyeprocess eye-tracking dataset", overwrite = FALSE, screen_distance_m = NULL, screen_size_m = NULL )
x |
An 'eye_dataset'. |
path |
BIDS dataset root. |
task |
Task label. |
dataset_name |
Dataset name for 'dataset_description.json'. |
overwrite |
Whether to replace existing files. |
screen_distance_m |
Optional screen distance in metres. |
screen_size_m |
Optional two-element screen size in metres. |
A manifest of written files.
Export a pipeline manifest and optional run status
export_eye_pipeline(x, path)export_eye_pipeline(x, path)
x |
Pipeline or run. |
path |
CSV path. |
An R object containing a pipeline manifest and optional run status. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
This is a compact interoperability representation inspired by W3C PROV. It is not asserted to be a complete PROV-O serialization; consumers requiring full conformance should validate/transform it externally.
export_prov_json(x, path)export_prov_json(x, path)
x |
Provenance graph. |
path |
Output JSON path. |
An R object containing lightweight PROV-oriented JSON. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Export minimal RO-Crate 1.3 metadata
export_ro_crate_metadata( path = "ro-crate-metadata.json", name = "eyeprocess analysis", description = "Reproducible eyeprocess analysis crate", files = NULL, creator = NULL, license = NULL, doi = NULL )export_ro_crate_metadata( path = "ro-crate-metadata.json", name = "eyeprocess analysis", description = "Reproducible eyeprocess analysis crate", files = NULL, creator = NULL, license = NULL, doi = NULL )
path |
Output 'ro-crate-metadata.json' path. |
name |
Crate/dataset name. |
description |
Description. |
files |
Optional files to include as File entities. |
creator |
Optional creator name. |
license |
Optional license URL or identifier. |
doi |
Optional DOI for the software/data product. |
An R object containing minimal RO-Crate 1.3 metadata. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Export a validation evidence bundle
export_validation_bundle(x, directory, overwrite = FALSE, include_rds = TRUE)export_validation_bundle(x, directory, overwrite = FALSE, include_rds = TRUE)
x |
Validation bundle. |
directory |
Output directory. |
overwrite |
Allow writing into a non-empty target directory. |
include_rds |
Save full R objects as RDS. |
An object of class "eye_validation_export", stored as a named list, with components "directory", "files", "manifest". It contains a validation evidence bundle and associated metadata or diagnostics needed to interpret the result.
Convert common external eye-tracking objects
as_eyeprocess_eyetools(x, mapping = NULL, ...) as_eyeprocess_eyetrackingr(x, mapping = NULL, ...) as_eyeprocess_gazer(x, mapping = NULL, ...) as_eyeprocess_eyeris(x, mapping = NULL, ...) as_eyeprocess_pupillometryr(x, mapping = NULL, ...)as_eyeprocess_eyetools(x, mapping = NULL, ...) as_eyeprocess_eyetrackingr(x, mapping = NULL, ...) as_eyeprocess_gazer(x, mapping = NULL, ...) as_eyeprocess_eyeris(x, mapping = NULL, ...) as_eyeprocess_pupillometryr(x, mapping = NULL, ...)
x |
External object coercible to a data frame. |
mapping |
Optional 'eye_mapping'. |
... |
Passed to 'read_eye_generic()'. |
An 'eye_dataset'.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
external_model_engines()external_model_engines()
The documented eyeprocess object, data frame, plot, report path, or adapter result.
External validation on a completely held-out dataset
external_validate_irt( train_data, external_data, fitter, predictor, scorer, label = "external" )external_validate_irt( train_data, external_data, fitter, predictor, scorer, label = "external" )
train_data |
Value supplied to 'train_data'; see Details for its model-specific role. |
external_data |
Value supplied to 'external_data'; see Details for its model-specific role. |
fitter |
Model-fitting function. |
predictor |
Prediction function. |
scorer |
Function that scores predictions. |
label |
Value supplied to 'label'; see Details for its model-specific role. |
A logical value or vector indicating external validation on a completely held-out dataset.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
extract_diffusion_parameters(object, variables = c("beta_drift", "beta_boundary", "beta_nondecision", "beta_starting", "person_drift", "item_difficulty", "boundary", "nondecision", "starting", "contaminant_probability"))extract_diffusion_parameters(object, variables = c("beta_drift", "beta_boundary", "beta_nondecision", "beta_starting", "person_drift", "item_difficulty", "boundary", "nondecision", "starting", "contaminant_probability"))
object |
Diffusion fit. |
variables |
Optional variable prefixes. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
extract_functional_pupil_parameters(x, pattern = NULL, confidence = 0.95)extract_functional_pupil_parameters(x, pattern = NULL, confidence = 0.95)
x |
Functional pupil fit. |
pattern |
Optional parameter regex. |
confidence |
Credible/confidence interval level. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Extract canonical parameter truth from simulated data
extract_parameter_truth(simulation)extract_parameter_truth(simulation)
simulation |
Simulation object containing 'truth', 'parameters', or a 'truth' attribute. |
Data frame with 'parameter' and 'truth'.
Extract standardized sliding-window process features
extract_process_windows( data, person = "person_id", trial = "trial_id", time = "time_ms", spec = process_window_spec(), align_time = NULL, pupil = "pupil_bc", pupil_tonic = "pupil_tonic", pupil_phasic = "pupil_phasic", gaze_x = "x", gaze_y = "y", aoi = "aoi", valid_gaze = "valid_gaze_prop", valid_pupil = "valid_pupil_prop", blink = "blink", trackloss = "trackloss" )extract_process_windows( data, person = "person_id", trial = "trial_id", time = "time_ms", spec = process_window_spec(), align_time = NULL, pupil = "pupil_bc", pupil_tonic = "pupil_tonic", pupil_phasic = "pupil_phasic", gaze_x = "x", gaze_y = "y", aoi = "aoi", valid_gaze = "valid_gaze_prop", valid_pupil = "valid_pupil_prop", blink = "blink", trackloss = "trackloss" )
data |
Sample-level eye-tracking/pupil data. |
person, trial, time
|
Identifier/time columns. |
spec |
Window specification. |
align_time |
Optional column containing the alignment event time in the same units as 'time'. Required for response/custom alignment unless 'time' is already relative to the desired origin. |
pupil |
Optional pupil signal column. |
pupil_tonic, pupil_phasic
|
Optional decomposed pupil columns. |
gaze_x, gaze_y
|
Optional gaze coordinates. |
aoi |
Optional AOI-state column. |
valid_gaze, valid_pupil
|
Optional validity columns/indicators. |
blink, trackloss
|
Optional blink/trackloss indicators. |
An 'eye_process_windows' object.
Construct a governed analysis pipeline
eye_analysis_pipeline( ..., spec = eye_analysis_spec(), name = "eye_analysis", strict = TRUE )eye_analysis_pipeline( ..., spec = eye_analysis_spec(), name = "eye_analysis", strict = TRUE )
... |
'eye_pipeline_step' objects. |
spec |
Optional 'eye_analysis_spec'. |
name |
Pipeline label. |
strict |
If 'TRUE', undeclared dependencies are errors. |
An object of class "eye_analysis_pipeline", stored as a named list, with components "name", "steps", "spec", "strict", "created_at", "status". It contains a governed analysis pipeline and associated metadata or diagnostics needed to interpret the result.
Define explicit analysis decisions for an eyeprocess workflow
eye_analysis_spec( blink_correction = NULL, pupil_baseline = NULL, fixation_algorithm = NULL, aoi_rule = NULL, exclusions = NULL, model = NULL, sensitivity = NULL, ... )eye_analysis_spec( blink_correction = NULL, pupil_baseline = NULL, fixation_algorithm = NULL, aoi_rule = NULL, exclusions = NULL, model = NULL, sensitivity = NULL, ... )
blink_correction |
Named blink-correction rule or 'NULL'. |
pupil_baseline |
Numeric baseline window or 'NULL'. |
fixation_algorithm |
Named fixation algorithm or 'NULL'. |
aoi_rule |
Named AOI assignment rule or 'NULL'. |
exclusions |
Explicit exclusion specification or 'NULL'. |
model |
Explicit model specification or 'NULL'. |
sensitivity |
Sensitivity specification or 'NULL'. |
... |
Additional named decisions. |
An 'eye_analysis_spec' object.
Inventory the public eyeprocess API
eye_api_inventory(package = "eyeprocess", lifecycle = eye_api_lifecycle())eye_api_inventory(package = "eyeprocess", lifecycle = eye_api_lifecycle())
package |
Package name or namespace environment. |
lifecycle |
Lifecycle registry. Defaults to the packaged 0.9 registry from 'eye_api_lifecycle()'. |
Data frame of exported symbols and lifecycle metadata.
Normalize or create an API lifecycle registry
eye_api_lifecycle(registry = NULL)eye_api_lifecycle(registry = NULL)
registry |
Optional data frame with at least 'name' and 'status'. 'NULL' loads the packaged 0.9 lifecycle registry. |
'eye_api_lifecycle' data frame.
Lifecycle recommendation for API review
eye_api_recommendation(audit)eye_api_recommendation(audit)
audit |
'eye_api_audit' object. |
A data frame containing lifecycle recommendation for API review. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Lookup lifecycle status for one or more APIs
eye_api_status(name, registry = eye_api_lifecycle())eye_api_status(name, registry = eye_api_lifecycle())
name |
API names. |
registry |
Lifecycle registry. |
A data frame containing lookup lifecycle status for one or more APIs. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Return superseded/deprecated compatibility interfaces
eye_api_superseded(registry = eye_api_lifecycle())eye_api_superseded(registry = eye_api_lifecycle())
registry |
Lifecycle registry. |
A tabular R object containing return superseded/deprecated compatibility interfaces; rows represent analysis units and columns contain the returned quantities.
Define a computational benchmark design
eye_benchmark_design( n_obs = c(10000, 1e+05, 1e+06), repetitions = 3L, label = "eyeprocess_scaling" )eye_benchmark_design( n_obs = c(10000, 1e+05, 1e+06), repetitions = 3L, label = "eyeprocess_scaling" )
n_obs |
Observation counts. |
repetitions |
Repetitions per size. |
label |
Benchmark label. |
A tabular R object containing define a computational benchmark design; rows represent analysis units and columns contain the returned quantities.
Create a machine-readable research decision manifest
eye_decision_manifest( sampling = list(), validity = list(), fixation = list(), pupil = list(), aoi = list(), model = list(), sensitivity = list(), exclusions = list(), provenance = list(), notes = NULL, ... )eye_decision_manifest( sampling = list(), validity = list(), fixation = list(), pupil = list(), aoi = list(), model = list(), sensitivity = list(), exclusions = list(), provenance = list(), notes = NULL, ... )
sampling |
Sampling decisions. |
validity |
Validity/missingness decisions. |
fixation |
Fixation construction decisions. |
pupil |
Pupil preprocessing decisions. |
aoi |
AOI assignment decisions. |
model |
Statistical/psychometric model decisions. |
sensitivity |
Sensitivity-analysis decisions. |
exclusions |
Exclusion rules. |
provenance |
Optional provenance fields. |
notes |
Notes. |
... |
Additional named decision domains. |
An object of class "eye_decision_manifest", stored as a named list, with components "domains", "notes", "created_at", "schema_version", "status", "caveat". It contains a machine-readable research decision manifest and associated metadata or diagnostics needed to interpret the result.
Render pipeline dependencies as Graphviz DOT
eye_pipeline_dot(x)eye_pipeline_dot(x)
x |
Pipeline. |
A character value or vector containing render pipeline dependencies as Graphviz DOT.
Return pipeline vertices and dependency edges
eye_pipeline_graph(x)eye_pipeline_graph(x)
x |
Pipeline. |
A named list with components "vertices", "edges", containing return pipeline vertices and dependency edges and associated metadata or diagnostics.
Machine-readable pipeline manifest
eye_pipeline_manifest(x)eye_pipeline_manifest(x)
x |
Pipeline. |
A data frame containing machine-readable pipeline manifest. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Render pipeline dependencies as Mermaid flowchart text
eye_pipeline_mermaid(x)eye_pipeline_mermaid(x)
x |
Pipeline. |
A character value or vector containing render pipeline dependencies as Mermaid flowchart text.
Define a governed pipeline step
eye_pipeline_step( name, fun, requires = character(), optional = FALSE, description = NULL, decision = NULL )eye_pipeline_step( name, fun, requires = character(), optional = FALSE, description = NULL, decision = NULL )
name |
Unique step name. |
fun |
Function executed by the step. |
requires |
Names of upstream steps. |
optional |
Whether an error may be retained without stopping the pipeline. |
description |
Human-readable description. |
decision |
Optional decision label linking the step to an analysis specification. |
An 'eye_pipeline_step' object.
Measurement-intelligence plotting and result infrastructure. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
eye_plot_spec(type = "default", title = NULL, xlab = NULL, ylab = NULL, caption = NULL, show_uncertainty = TRUE, show_raw = TRUE, facet_by = NULL, label_items = FALSE, interactive = FALSE) plot_diagnostics(x, ...) plot_evidence(x, ...) plot_sensitivity(x, ...) autoplot_eyeprocess(object, ...)eye_plot_spec(type = "default", title = NULL, xlab = NULL, ylab = NULL, caption = NULL, show_uncertainty = TRUE, show_raw = TRUE, facet_by = NULL, label_items = FALSE, interactive = FALSE) plot_diagnostics(x, ...) plot_evidence(x, ...) plot_sensitivity(x, ...) autoplot_eyeprocess(object, ...)
type |
Argument controlling 'type'; see the function usage and returned audit metadata. |
title |
Argument controlling 'title'; see the function usage and returned audit metadata. |
xlab |
Argument controlling 'xlab'; see the function usage and returned audit metadata. |
ylab |
Argument controlling 'ylab'; see the function usage and returned audit metadata. |
caption |
Argument controlling 'caption'; see the function usage and returned audit metadata. |
show_uncertainty |
Argument controlling 'show_uncertainty'; see the function usage and returned audit metadata. |
show_raw |
Argument controlling 'show_raw'; see the function usage and returned audit metadata. |
facet_by |
Argument controlling 'facet_by'; see the function usage and returned audit metadata. |
label_items |
Argument controlling 'label_items'; see the function usage and returned audit metadata. |
interactive |
Argument controlling 'interactive'; see the function usage and returned audit metadata. |
x |
Input object or data structure appropriate for the selected analysis. |
... |
Additional arguments passed to the underlying method or plotting function. |
object |
Input object or data structure appropriate for the selected analysis. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Construct a lightweight provenance graph
eye_prov_graph( nodes, edges = data.frame(from = character(), to = character(), relation = character()), metadata = list() )eye_prov_graph( nodes, edges = data.frame(from = character(), to = character(), relation = character()), metadata = list() )
nodes |
Node table. |
edges |
Edge table. |
metadata |
Optional metadata. |
A named list with components "nodes", "edges", "metadata", containing a lightweight provenance graph and associated metadata or diagnostics.
Construct a reproducibility fingerprint
eye_reproducibility_fingerprint( data = NULL, analysis_spec = NULL, model_spec = NULL, decisions = NULL, result = NULL, files = NULL, seeds = NULL, label = "eyeprocess_analysis" )eye_reproducibility_fingerprint( data = NULL, analysis_spec = NULL, model_spec = NULL, decisions = NULL, result = NULL, files = NULL, seeds = NULL, label = "eyeprocess_analysis" )
data |
Input data or hashable object. |
analysis_spec |
Analysis specification. |
model_spec |
Model specification. |
decisions |
Decision manifest. |
result |
Optional result object. |
files |
Optional input file paths. |
seeds |
Optional seeds. |
label |
Fingerprint label. |
A named list with components "schema_version", "label", "eyeprocess_version", "data_hash", "analysis_spec_hash", "model_spec_hash", "decisions_hash", "result_hash", "file_manifest", "seeds", "environment", containing a reproducibility fingerprint and associated metadata or diagnostics.
Create a session-level provenance manifest
eye_session_manifest( data = NULL, files = NULL, adapter = NA_character_, decisions = NULL, pipeline = NULL, seeds = NULL, notes = NULL )eye_session_manifest( data = NULL, files = NULL, adapter = NA_character_, decisions = NULL, pipeline = NULL, seeds = NULL, notes = NULL )
data |
Optional data object. |
files |
Optional input file paths. |
adapter |
Adapter/importer identifier. |
decisions |
Optional decision manifest. |
pipeline |
Optional pipeline or run object. |
seeds |
Optional named random seeds. |
notes |
Optional notes. |
A named list with components "created_utc", "eyeprocess_version", "data_hash", "files", "adapter", "decisions_hash", "pipeline_hash", "seeds", "environment", "notes", containing a session-level provenance manifest and associated metadata or diagnostics.
Specify disk-backed eyeprocess storage
eye_storage_spec( path, format = c("rds", "parquet", "arrow_dataset"), tables = canonical_table_names(), partitioning = NULL, compression = "zstd" )eye_storage_spec( path, format = c("rds", "parquet", "arrow_dataset"), tables = canonical_table_names(), partitioning = NULL, compression = "zstd" )
path |
Storage directory or RDS file. |
format |
One of '"rds"', '"parquet"', or '"arrow_dataset"'. |
tables |
Canonical tables to store. |
partitioning |
Optional Arrow partition columns. |
compression |
Parquet compression codec. For writes, the default '"zstd"' is preferred; when the argument is omitted and that codec is unavailable, storage falls back to '"snappy"' and then '"uncompressed"'. An explicitly requested unavailable codec errors. |
An 'eye_storage_spec' object.
Audit preservation of monocular/binocular stream semantics
eye_stream_fidelity_audit( source, roundtrip, source_eye = "eye", roundtrip_eye = source_eye, key = NULL )eye_stream_fidelity_audit( source, roundtrip, source_eye = "eye", roundtrip_eye = source_eye, key = NULL )
source |
Original/source representation. |
roundtrip |
Round-tripped or comparison representation. |
source_eye |
Source recorded-eye column. |
roundtrip_eye |
Round-tripped recorded-eye column. |
key |
Column or columns used to align records. |
An object of class "eye_stream_fidelity", stored as a named list, with components "status", "matched_n", "source_streams", "roundtrip_streams", "confusion". It contains preservation of monocular/binocular stream semantics and associated metadata or diagnostics needed to interpret the result.
This function does not execute or silently translate arbitrary closures into a targets pipeline. It provides the dependency contract required to build an explicit '_targets.R' file.
eye_targets_manifest(x)eye_targets_manifest(x)
x |
Pipeline. |
A data frame containing a targets-compatible dependency manifest. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
eyeprocess_api_version()eyeprocess_api_version()
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
eyeprocess_benchmark_study(path = NULL)eyeprocess_benchmark_study(path = NULL)
path |
Optional alternative benchmark root. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Enumerate latent attribute profiles
eyeprocess_cdm_attribute_profiles( n_attributes, attribute_names = paste0("A", seq_len(n_attributes)) )eyeprocess_cdm_attribute_profiles( n_attributes, attribute_names = paste0("A", seq_len(n_attributes)) )
n_attributes |
Number of cognitive-diagnosis attributes. |
attribute_names |
Optional names for the cognitive-diagnosis attributes. |
An R object containing enumerate latent attribute profiles. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarise CDM classification uncertainty from profile probabilities
eyeprocess_cdm_classification_uncertainty(profile_probabilities)eyeprocess_cdm_classification_uncertainty(profile_probabilities)
profile_probabilities |
Posterior attribute-profile probabilities. |
A data frame containing cDM classification uncertainty from profile probabilities. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compute deterministic DINA ideal responses from a Q-matrix
eyeprocess_cdm_dina_ideal_response(Q, profiles)eyeprocess_cdm_dina_ideal_response(Q, profiles)
Q |
Binary item-by-attribute Q-matrix. |
profiles |
Attribute mastery profiles, with rows representing profiles. |
A vector or matrix containing deterministic DINA ideal responses from a Q-matrix, with shape determined by the supplied analysis units.
DINA response probabilities from slip and guess parameters
eyeprocess_cdm_dina_probability(ideal_response, slip = 0.1, guess = 0.2)eyeprocess_cdm_dina_probability(ideal_response, slip = 0.1, guess = 0.2)
ideal_response |
Ideal-response indicator or matrix implied by the cognitive-diagnosis model. |
slip |
DINA slip parameter or vector of slip parameters. |
guess |
DINA guessing parameter or vector of guessing parameters. |
An object of class "matrix", stored as an R object, containing dINA response probabilities from slip and guess parameters and associated metadata needed to interpret the result.
Audit a cognitive-diagnosis Q-matrix
eyeprocess_cdm_qmatrix_audit( Q, item_ids = rownames(Q), attribute_names = colnames(Q) )eyeprocess_cdm_qmatrix_audit( Q, item_ids = rownames(Q), attribute_names = colnames(Q) )
Q |
Binary item-by-attribute Q-matrix. |
item_ids |
Optional item identifiers. |
attribute_names |
Optional names for the cognitive-diagnosis attributes. |
An object of class "eye_cdm_qmatrix_audit", stored as a named list, with components "item", "attribute", "duplicate_rows", "complete_identity_block". It contains a cognitive-diagnosis Q-matrix and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
eyeprocess_deprecation(old, replacement, since, remove_after, reason = "")eyeprocess_deprecation(old, replacement, since, remove_after, reason = "")
old |
Deprecated symbol. |
replacement |
Replacement symbol. |
since |
Version where deprecation started. |
remove_after |
Earliest removal version. |
reason |
Reason. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
2PL item-response probability
eyeprocess_irt_2pl_probability(theta, a = 1, b = 0, D = 1)eyeprocess_irt_2pl_probability(theta, a = 1, b = 0, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
a |
Item discrimination parameter or vector. |
b |
Item difficulty or location parameter or vector. |
D |
Logistic scaling constant. |
A numeric value or vector containing 2PL item-response probability.
3PL item-response probability
eyeprocess_irt_3pl_probability(theta, a = 1, b = 0, c = 0.2, D = 1)eyeprocess_irt_3pl_probability(theta, a = 1, b = 0, c = 0.2, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
a |
Item discrimination parameter or vector. |
b |
Item difficulty or location parameter or vector. |
c |
Lower-asymptote parameter or vector. |
D |
Logistic scaling constant. |
A numeric value or vector containing 3PL item-response probability.
4PL item-response probability
eyeprocess_irt_4pl_probability(theta, a = 1, b = 0, c = 0, d = 1, D = 1)eyeprocess_irt_4pl_probability(theta, a = 1, b = 0, c = 0, d = 1, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
a |
Item discrimination parameter or vector. |
b |
Item difficulty or location parameter or vector. |
c |
Lower-asymptote parameter or vector. |
d |
Upper-asymptote parameter or vector. |
D |
Logistic scaling constant. |
A numeric value or vector containing 4PL item-response probability.
Create an auditable adaptive-testing trace
eyeprocess_irt_adaptive_trace( item_id, theta_before, theta_after, se_after, information, response = NA_real_ )eyeprocess_irt_adaptive_trace( item_id, theta_before, theta_after, se_after, information, response = NA_real_ )
item_id |
Item identifier or vector of item identifiers. |
theta_before |
Latent-trait estimate before item administration. |
theta_after |
Latent-trait estimate after item administration. |
se_after |
Conditional standard error after item administration. |
information |
Item or test information value or vector. |
response |
Observed item response or response variable. |
An object of class "eye_irt_adaptive_trace", "data.frame", stored as a data frame, containing an auditable adaptive-testing trace and associated metadata needed to interpret the result.
Audit candidate anchor items using supplied DIF evidence
eyeprocess_irt_anchor_audit( items, dif = NULL, max_abs_effect = 0.1, min_information = NULL )eyeprocess_irt_anchor_audit( items, dif = NULL, max_abs_effect = 0.1, min_information = NULL )
items |
Item-parameter data frame or item collection. |
dif |
Differential-item-functioning evidence or summary. |
max_abs_effect |
Maximum permitted absolute effect for an anchor candidate. |
min_information |
Minimum required item information. |
A data frame containing candidate anchor items using supplied DIF evidence. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Iteratively remove anchors exceeding a supplied effect threshold
eyeprocess_irt_anchor_purification( items, effect_fun, initial = items$item_id, threshold = 0.1, max_iter = 10L )eyeprocess_irt_anchor_purification( items, effect_fun, initial = items$item_id, threshold = 0.1, max_iter = 10L )
items |
Item-parameter data frame or item collection. |
effect_fun |
Function that computes the anchor-screening effect. |
initial |
Initial value, state, or anchor set. |
threshold |
Decision or diagnostic threshold. |
max_iter |
Maximum number of iterations. |
An object of class "eye_irt_anchor_purification", stored as a named list, with components "anchors", "history", "threshold". It contains iteratively remove anchors exceeding a supplied effect threshold and associated metadata or diagnostics needed to interpret the result.
Apply linear IRT scale-linking coefficients
eyeprocess_irt_apply_link(items, link)eyeprocess_irt_apply_link(items, link)
items |
Item-parameter data frame or item collection. |
link |
IRT scale-linking coefficients or linking object. |
An R object containing linear IRT scale-linking coefficients. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Audit item-bank information coverage across a theta region
eyeprocess_irt_bank_coverage( items, theta = seq(-4, 4, length.out = 161), target_information = 5, target = c(-2, 2) )eyeprocess_irt_bank_coverage( items, theta = seq(-4, 4, length.out = 161), target_information = 5, target = c(-2, 2) )
items |
Item-parameter data frame or item collection. |
theta |
Latent-trait value or vector of latent-trait values. |
target_information |
Target test-information level. |
target |
Target level, distribution, or criterion. |
An object of class "eye_irt_bank_coverage", stored as a named list, with components "curve", "target", "target_information", "fraction_target_met", "minimum_information", "maximum_sem", "gaps". It contains item-bank information coverage across a theta region and associated metadata or diagnostics needed to interpret the result.
Audit category probability functions
eyeprocess_irt_category_function_audit(probabilities, tolerance = 1e-08)eyeprocess_irt_category_function_audit(probabilities, tolerance = 1e-08)
probabilities |
Probability matrix or vector, with dimensions appropriate to the model. |
tolerance |
Numerical or decision tolerance. |
An object of class "eye_irt_category_audit", stored as a named list, with components "valid_bounds", "rows_sum_to_one", "max_sum_error", "min_probability", "max_probability". It contains category probability functions and associated metadata or diagnostics needed to interpret the result.
Summarise decision precision at one or more theta cut scores
eyeprocess_irt_classification_precision( theta_estimate, standard_error, cut_score = 0, confidence = 0.95 )eyeprocess_irt_classification_precision( theta_estimate, standard_error, cut_score = 0, confidence = 0.95 )
theta_estimate |
Estimated latent-trait values. |
standard_error |
Standard errors corresponding to the estimates. |
cut_score |
Latent-scale classification cut score. |
confidence |
Requested confidence level. |
A tabular R object containing decision precision at one or more theta cut scores; rows represent analysis units and columns contain the returned quantities.
Conditional standard error from information
eyeprocess_irt_conditional_sem(information)eyeprocess_irt_conditional_sem(information)
information |
Item or test information value or vector. |
A logical value or vector indicating conditional standard error from information.
Audit content balance in an administered adaptive form
eyeprocess_irt_content_balance_audit(administered, item_bank, target = NULL)eyeprocess_irt_content_balance_audit(administered, item_bank, target = NULL)
administered |
Identifiers or records for administered items. |
item_bank |
Item bank or item-bank data frame. |
target |
Target level, distribution, or criterion. |
A data frame containing content balance in an administered adaptive form. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Summarise parameter drift across acquisition devices
eyeprocess_irt_device_drift( parameters, item_id = "item_id", device = "device", parameter = "b" )eyeprocess_irt_device_drift( parameters, item_id = "item_id", device = "device", parameter = "b" )
parameters |
Item-parameter data frame or parameter estimates. |
item_id |
Item identifier or vector of item identifiers. |
device |
Device identifier or grouping variable. |
parameter |
Name of the parameter to compare. |
A tabular R object containing parameter drift across acquisition devices; rows represent analysis units and columns contain the returned quantities.
Differential item functioning effect curve from two parameter sets
eyeprocess_irt_dif_effect_curve( reference_item, focal_item, theta = seq(-4, 4, length.out = 81) )eyeprocess_irt_dif_effect_curve( reference_item, focal_item, theta = seq(-4, 4, length.out = 81) )
reference_item |
Reference-group item parameters. |
focal_item |
Focal-group item parameters. |
theta |
Latent-trait value or vector of latent-trait values. |
An object of class "eye_irt_dif_curve", "data.frame", stored as a data frame, containing differential item functioning effect curve from two parameter sets and associated metadata needed to interpret the result.
Differential test functioning effect curve
eyeprocess_irt_dtf_curve(reference, focal, theta = seq(-4, 4, length.out = 81))eyeprocess_irt_dtf_curve(reference, focal, theta = seq(-4, 4, length.out = 81))
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
theta |
Latent-trait value or vector of latent-trait values. |
An object of class "eye_irt_dtf_curve", "data.frame", stored as a data frame, containing differential test functioning effect curve and associated metadata needed to interpret the result.
EAP score for dichotomous IRT item parameters
eyeprocess_irt_eap_score( response, items, theta_grid = seq(-4, 4, length.out = 81), prior_mean = 0, prior_sd = 1, D = 1 )eyeprocess_irt_eap_score( response, items, theta_grid = seq(-4, 4, length.out = 81), prior_mean = 0, prior_sd = 1, D = 1 )
response |
Observed item response or response variable. |
items |
Item-parameter data frame or item collection. |
theta_grid |
Grid of latent-trait values used for numerical scoring or integration. |
prior_mean |
Mean of the normal latent-trait prior. |
prior_sd |
Standard deviation of the normal latent-trait prior. |
D |
Logistic scaling constant. |
An object of class "eye_irt_score", stored as a named list, with components "estimate", "se", "theta", "posterior", "method". It contains eAP score for dichotomous IRT item parameters and associated metadata or diagnostics needed to interpret the result.
Build an external-engine capability and availability table
eyeprocess_irt_engine_evidence_table()eyeprocess_irt_engine_evidence_table()
A data frame containing an external-engine capability and availability table. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Registry of specialized external IRT engines
eyeprocess_irt_engine_registry()eyeprocess_irt_engine_registry()
An object of class "eye_irt_engine_registry", "data.frame", stored as a data frame, containing registry of specialized external IRT engines and associated metadata needed to interpret the result.
Query an external IRT engine
eyeprocess_irt_engine_status(engine)eyeprocess_irt_engine_status(engine)
engine |
Requested estimation or analysis engine. |
An object of class "eye_irt_engine_registry", "data.frame", stored as a data frame, containing query an external IRT engine and associated metadata needed to interpret the result.
Expected item score
eyeprocess_irt_expected_score( theta, family = c("2pl", "3pl", "4pl", "grm", "gpcm", "nominal"), ... )eyeprocess_irt_expected_score( theta, family = c("2pl", "3pl", "4pl", "grm", "gpcm", "nominal"), ... )
theta |
Latent-trait value or vector of latent-trait values. |
family |
IRT response family or model family. |
... |
Additional arguments passed to the selected method or external engine. |
A numeric value or vector containing expected item score.
Summarise item exposure rates
eyeprocess_irt_exposure_summary( administered, item_bank_ids = unique(administered) )eyeprocess_irt_exposure_summary( administered, item_bank_ids = unique(administered) )
administered |
Identifiers or records for administered items. |
item_bank_ids |
Complete set of item identifiers in the bank. |
A data frame containing item exposure rates. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Audit extreme response scores without assigning behavioral labels
eyeprocess_irt_extreme_score_audit( responses, lower_fraction = 0.02, upper_fraction = 0.98 )eyeprocess_irt_extreme_score_audit( responses, lower_fraction = 0.02, upper_fraction = 0.98 )
responses |
Response matrix or response data. |
lower_fraction |
Lower extreme-score fraction. |
upper_fraction |
Upper extreme-score fraction. |
A data frame containing extreme response scores without assigning behavioral labels. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Build an integrated IRT diagnostic dashboard object
eyeprocess_irt_fit_dashboard( item_fit = NULL, person_fit = NULL, q3 = NULL, parameter_audit = NULL, identification = NULL )eyeprocess_irt_fit_dashboard( item_fit = NULL, person_fit = NULL, q3 = NULL, parameter_audit = NULL, identification = NULL )
item_fit |
Item-fit diagnostic object or table. |
person_fit |
Person-fit diagnostic object or table. |
q3 |
Q3 residual-correlation matrix or summary. |
parameter_audit |
Item-parameter plausibility audit. |
identification |
Identification specification or identification audit. |
An object of class "eye_irt_fit_dashboard", stored as a named list, with components "components", "present", "n_components", "interpretation". It contains an integrated IRT diagnostic dashboard object and associated metadata or diagnostics needed to interpret the result.
Summarise DIF/DTF curve magnitude
eyeprocess_irt_functioning_effect_summary(curve)eyeprocess_irt_functioning_effect_summary(curve)
curve |
Curve data to summarize or inspect. |
A named list with components "max_abs", "mean_abs", "signed_area", containing dIF/DTF curve magnitude and associated metadata or diagnostics.
Generalized partial-credit category probabilities
eyeprocess_irt_gpcm_probability(theta, a = 1, steps, D = 1)eyeprocess_irt_gpcm_probability(theta, a = 1, steps, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
a |
Item discrimination parameter or vector. |
steps |
Step parameters for a generalized partial-credit model. |
D |
Logistic scaling constant. |
A logical value or vector indicating generalized partial-credit category probabilities.
Graded-response category probabilities
eyeprocess_irt_grm_probability(theta, a = 1, thresholds, D = 1)eyeprocess_irt_grm_probability(theta, a = 1, thresholds, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
a |
Item discrimination parameter or vector. |
thresholds |
Ordered response-category thresholds. |
D |
Logistic scaling constant. |
A logical value or vector indicating graded-response category probabilities.
Haebara item-characteristic-curve linking
eyeprocess_irt_haebara_link( reference, focal, anchors = NULL, theta = seq(-4, 4, length.out = 81), weights = NULL, start = c(A = 1, B = 0) )eyeprocess_irt_haebara_link( reference, focal, anchors = NULL, theta = seq(-4, 4, length.out = 81), weights = NULL, start = c(A = 1, B = 0) )
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
anchors |
Anchor-item identifiers. |
theta |
Latent-trait value or vector of latent-trait values. |
weights |
Optional numerical weights. |
start |
Starting values for numerical optimization. |
An object of class "eye_irt_link", stored as a named list, with components "A", "B", "method", "anchors", "objective", "convergence". It contains haebara item-characteristic-curve linking and associated metadata or diagnostics needed to interpret the result.
Audit IRT scale/location identification
eyeprocess_irt_identification_audit( spec, constraints = list(), n_items = NULL, n_persons = NULL )eyeprocess_irt_identification_audit( spec, constraints = list(), n_items = NULL, n_persons = NULL )
spec |
Model, validation, or analysis specification object. |
constraints |
Identification or model constraints. |
n_items |
Number of items. |
n_persons |
Number of persons. |
An object of class "eye_irt_identification_audit", stored as a named list, with components "spec", "location_identified", "scale_identified", "anchors", "warnings", "valid", "n_items", "n_persons". It contains iRT scale/location identification and associated metadata or diagnostics needed to interpret the result.
These statistics summarize response-model residual behavior. They are not labels for motivation, misconduct, diagnosis, or respondent intent.
eyeprocess_irt_infit_outfit( observed, expected, by = c("item", "person"), min_variance = 1e-08 )eyeprocess_irt_infit_outfit( observed, expected, by = c("item", "person"), min_variance = 1e-08 )
observed |
Observed responses or observed values. |
expected |
Model-expected probabilities or expected values. |
by |
Grouping variables or aggregation level. |
min_variance |
Minimum variance used to stabilize residual calculations. |
A data frame containing residual-based Infit and Outfit summaries. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Area under an information curve
eyeprocess_irt_information_area(theta, information)eyeprocess_irt_information_area(theta, information)
theta |
Latent-trait value or vector of latent-trait values. |
information |
Item or test information value or vector. |
A numeric value or vector containing area under an information curve.
Information gain between two conditional standard errors
eyeprocess_irt_information_gain(se_before, se_after)eyeprocess_irt_information_gain(se_before, se_after)
se_before |
Conditional standard error before an item is administered. |
se_after |
Conditional standard error after item administration. |
A numeric value or vector containing information gain between two conditional standard errors.
Audit how well item information targets a theta distribution
eyeprocess_irt_information_targeting(items, theta, weights = NULL, D = 1)eyeprocess_irt_information_targeting(items, theta, weights = NULL, D = 1)
items |
Item-parameter data frame or item collection. |
theta |
Latent-trait value or vector of latent-trait values. |
weights |
Optional numerical weights. |
D |
Logistic scaling constant. |
An object of class "eye_irt_information_targeting", stored as a named list, with components "weighted_information", "weighted_sem", "curve", "weights". It contains how well item information targets a theta distribution and associated metadata or diagnostics needed to interpret the result.
Combine invariance evidence without converting it to a binary validity claim
eyeprocess_irt_invariance_evidence( anchor_audit = NULL, dif = NULL, dtf = NULL, linking = NULL, process_concordance = NULL )eyeprocess_irt_invariance_evidence( anchor_audit = NULL, dif = NULL, dtf = NULL, linking = NULL, process_concordance = NULL )
anchor_audit |
Anchor-item audit result. |
dif |
Differential-item-functioning evidence or summary. |
dtf |
Differential-test-functioning evidence or curve. |
linking |
Scale-linking evidence or result. |
process_concordance |
Process-DIF concordance evidence. |
An object of class "eye_irt_invariance_evidence", stored as a named list, with components "components", "present", "completeness", "interpretation". It contains combine invariance evidence without converting it to a binary validity claim and associated metadata or diagnostics needed to interpret the result.
Item bank object for adaptive design
eyeprocess_irt_item_bank(items, content = NULL, exposure_limit = 1)eyeprocess_irt_item_bank(items, content = NULL, exposure_limit = 1)
items |
Item-parameter data frame or item collection. |
content |
Item content/category metadata. |
exposure_limit |
Maximum permitted item exposure. |
An object of class "eye_irt_item_bank", stored as a named list, with components "items", "exposure_limit". It contains item bank object for adaptive design and associated metadata or diagnostics needed to interpret the result.
Compute item residual fit summaries from observed and predicted probabilities
eyeprocess_irt_item_fit_residuals( responses, probabilities, item_ids = colnames(responses) )eyeprocess_irt_item_fit_residuals( responses, probabilities, item_ids = colnames(responses) )
responses |
Response matrix or response data. |
probabilities |
Probability matrix or vector, with dimensions appropriate to the model. |
item_ids |
Optional item identifiers. |
An object of class "eye_irt_item_fit", "data.frame", stored as a data frame, containing item residual fit summaries from observed and predicted probabilities and associated metadata needed to interpret the result.
Compute item information for transparent IRT families
eyeprocess_irt_item_information( theta, family = c("2pl", "3pl", "4pl", "grm", "gpcm", "nominal"), ..., D = 1 )eyeprocess_irt_item_information( theta, family = c("2pl", "3pl", "4pl", "grm", "gpcm", "nominal"), ..., D = 1 )
theta |
Latent-trait value or vector of latent-trait values. |
family |
IRT response family or model family. |
... |
Additional arguments passed to the selected method or external engine. |
D |
Logistic scaling constant. |
A numeric value or vector containing item information for transparent IRT families.
Select the most informative eligible item at a theta estimate
eyeprocess_irt_item_selection( bank, theta, administered = character(), exposure = NULL, content_required = NULL, D = 1 )eyeprocess_irt_item_selection( bank, theta, administered = character(), exposure = NULL, content_required = NULL, D = 1 )
bank |
Validated item-bank object. |
theta |
Latent-trait value or vector of latent-trait values. |
administered |
Identifiers or records for administered items. |
exposure |
Item exposure information used by the adaptive-selection rule. |
content_required |
Content constraints required for item selection. |
D |
Logistic scaling constant. |
An object of class "eye_irt_item_selection", stored as a named list, with components "selected", "information", "theta", "reason", "candidate_count". It contains the most informative eligible item at a theta estimate and associated metadata or diagnostics needed to interpret the result.
Build a latent-regression design matrix with explicit centering metadata
eyeprocess_irt_latent_regression_design(data, formula, center_numeric = TRUE)eyeprocess_irt_latent_regression_design(data, formula, center_numeric = TRUE)
data |
Input data frame, matrix, or compatible analysis object. |
formula |
Model formula. |
center_numeric |
Whether numeric predictors are centered. |
An object of class "eye_irt_latent_regression_design", stored as a named list, with components "matrix", "formula", "centers", "complete". It contains a latent-regression design matrix with explicit centering metadata and associated metadata or diagnostics needed to interpret the result.
Compare linking estimates across anchor subsets
eyeprocess_irt_link_stability( reference, focal, anchor_sets, method = c("mean-sigma", "mean-mean", "Stocking-Lord", "Haebara") )eyeprocess_irt_link_stability( reference, focal, anchor_sets, method = c("mean-sigma", "mean-mean", "Stocking-Lord", "Haebara") )
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
anchor_sets |
Value supplied for the anchor sets argument. |
method |
Scoring, linking, or analysis method. |
An object of class "eye_irt_link_stability", stored as a named list, with components "table", "sd_A", "sd_B", "method". It contains linking estimates across anchor subsets and associated metadata or diagnostics needed to interpret the result.
Extract high residual-dependence item pairs
eyeprocess_irt_local_dependence_pairs(q3, threshold = 0.2, absolute = TRUE)eyeprocess_irt_local_dependence_pairs(q3, threshold = 0.2, absolute = TRUE)
q3 |
Q3 residual-correlation matrix or summary. |
threshold |
Decision or diagnostic threshold. |
absolute |
Whether diagnostic thresholds apply to absolute values. |
A data frame containing high residual-dependence item pairs. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
MAP score for dichotomous IRT item parameters
eyeprocess_irt_map_score( response, items, bounds = c(-6, 6), prior_mean = 0, prior_sd = 1, D = 1 )eyeprocess_irt_map_score( response, items, bounds = c(-6, 6), prior_mean = 0, prior_sd = 1, D = 1 )
response |
Observed item response or response variable. |
items |
Item-parameter data frame or item collection. |
bounds |
Numerical lower and upper optimization bounds. |
prior_mean |
Mean of the normal latent-trait prior. |
prior_sd |
Standard deviation of the normal latent-trait prior. |
D |
Logistic scaling constant. |
An object of class "eye_irt_score", stored as a named list, with components "estimate", "objective", "method", "bounds". It contains mAP score for dichotomous IRT item parameters and associated metadata or diagnostics needed to interpret the result.
Marginal reliability from latent-score variance and conditional error variance
eyeprocess_irt_marginal_reliability(theta_estimate, se)eyeprocess_irt_marginal_reliability(theta_estimate, se)
theta_estimate |
Estimated latent-trait values. |
se |
Standard-error values. |
A numeric value or vector containing marginal reliability from latent-score variance and conditional error variance.
Mean-mean IRT linking coefficients
eyeprocess_irt_mean_mean_link(reference, focal, anchors = NULL)eyeprocess_irt_mean_mean_link(reference, focal, anchors = NULL)
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
anchors |
Anchor-item identifiers. |
An object of class "eye_irt_link", stored as a named list, with components "A", "B", "method", "anchors", "objective". It contains mean-mean IRT linking coefficients and associated metadata or diagnostics needed to interpret the result.
Mean-sigma IRT linking coefficients
eyeprocess_irt_mean_sigma_link(reference, focal, anchors = NULL)eyeprocess_irt_mean_sigma_link(reference, focal, anchors = NULL)
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
anchors |
Anchor-item identifiers. |
An object of class "eye_irt_link", stored as a named list, with components "A", "B", "method", "anchors", "objective". It contains mean-sigma IRT linking coefficients and associated metadata or diagnostics needed to interpret the result.
Summarise measurement precision across a theta region
eyeprocess_irt_measurement_precision_profile( theta, items, target = c(-2, 2), D = 1 )eyeprocess_irt_measurement_precision_profile( theta, items, target = c(-2, 2), D = 1 )
theta |
Latent-trait value or vector of latent-trait values. |
items |
Item-parameter data frame or item collection. |
target |
Target level, distribution, or criterion. |
D |
Logistic scaling constant. |
An object of class "eye_irt_precision_profile", stored as a named list, with components "curve", "target", "area", "min_information", "max_sem". It contains measurement precision across a theta region and associated metadata or diagnostics needed to interpret the result.
Audit missing-by-design structure in an IRT response matrix
eyeprocess_irt_missing_by_design_audit( responses, design = NULL, min_administered = 1L )eyeprocess_irt_missing_by_design_audit( responses, design = NULL, min_administered = 1L )
responses |
Response matrix or response data. |
design |
Validation or simulation design object. |
min_administered |
Minimum number of administered items required for a record. |
An object of class "eye_irt_missing_design_audit", stored as a named list, with components "n_persons", "n_items", "observed_fraction", "administered_per_person", "administered_per_item", "sparse_persons", "structural_missing", "unexpected_missing", "has_declared_design". It contains missing-by-design structure in an IRT response matrix and associated metadata or diagnostics needed to interpret the result.
Compare recovery under reference and misspecified scenarios
eyeprocess_irt_misspecification_metrics( reference_summary, misspecified_summary )eyeprocess_irt_misspecification_metrics( reference_summary, misspecified_summary )
reference_summary |
Reference-model validation summary. |
misspecified_summary |
Misspecified-model validation summary. |
An R object containing recovery under reference and misspecified scenarios. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Create a model-misspecification suite
eyeprocess_irt_misspecification_suite()eyeprocess_irt_misspecification_suite()
A data frame containing a model-misspecification suite. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Bounded ML score for dichotomous IRT item parameters
eyeprocess_irt_mle_score(response, items, bounds = c(-6, 6), D = 1)eyeprocess_irt_mle_score(response, items, bounds = c(-6, 6), D = 1)
response |
Observed item response or response variable. |
items |
Item-parameter data frame or item collection. |
bounds |
Numerical lower and upper optimization bounds. |
D |
Logistic scaling constant. |
An object of class "eye_irt_score", stored as a named list, with components "estimate", "objective", "method", "bounds", "boundary". It contains bounded ML score for dichotomous IRT item parameters and associated metadata or diagnostics needed to interpret the result.
Create a governed IRT model card
eyeprocess_irt_model_card( spec, engine_status = NULL, identification = NULL, fit_evidence = NULL, invariance = NULL, validation = NULL, intended_use = NULL, excluded_interpretations = c("diagnosis", "cheating inference", "mental-state inference") )eyeprocess_irt_model_card( spec, engine_status = NULL, identification = NULL, fit_evidence = NULL, invariance = NULL, validation = NULL, intended_use = NULL, excluded_interpretations = c("diagnosis", "cheating inference", "mental-state inference") )
spec |
Model, validation, or analysis specification object. |
engine_status |
Availability/status record for the selected estimation engine. |
identification |
Identification specification or identification audit. |
fit_evidence |
Model-fit evidence or diagnostics. |
invariance |
Measurement-invariance evidence or audit. |
validation |
Value supplied for the validation argument. |
intended_use |
Statement of the intended analytical use. |
excluded_interpretations |
Interpretations explicitly excluded by the model card. |
An object of class "eye_irt_model_card", stored as a named list, with components "specification", "engine_status", "identification", "fit_evidence", "invariance", "validation", "intended_use", "excluded_interpretations", "created", "hash". It contains a governed IRT model card and associated metadata or diagnostics needed to interpret the result.
Audit completeness of an IRT model card
eyeprocess_irt_model_card_audit(card)eyeprocess_irt_model_card_audit(card)
card |
Value supplied for the card argument. |
A data frame containing completeness of an IRT model card. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
This specification records a psychometric model contract. Estimation is delegated to explicit engines where required; the specification itself does not fit a model.
eyeprocess_irt_model_spec( family = c("rasch", "2pl", "3pl", "4pl", "grm", "gpcm", "nominal", "multidimensional", "testlet", "latent_regression", "cdm", "joint_rt"), dimensions = 1L, identification = c("theta_standard", "item_sum_zero", "anchor"), engine = c("native_math", "mirt", "TAM", "GDINA", "LNIRT", "eRm", "custom"), process_channels = character(), status = c("reference", "experimental", "gated"), notes = NULL )eyeprocess_irt_model_spec( family = c("rasch", "2pl", "3pl", "4pl", "grm", "gpcm", "nominal", "multidimensional", "testlet", "latent_regression", "cdm", "joint_rt"), dimensions = 1L, identification = c("theta_standard", "item_sum_zero", "anchor"), engine = c("native_math", "mirt", "TAM", "GDINA", "LNIRT", "eRm", "custom"), process_channels = character(), status = c("reference", "experimental", "gated"), notes = NULL )
family |
IRT response family or model family. |
dimensions |
Value supplied for the dimensions argument. |
identification |
Identification specification or identification audit. |
engine |
Requested estimation or analysis engine. |
process_channels |
Declared process-measure channels. |
status |
Evidence, model, or governance status. |
notes |
Value supplied for the notes argument. |
An object of class "eyeprocess_irt_model_spec", stored as a named list, with components "family", "dimensions", "identification", "engine", "process_channels", "status", "notes". It contains declare an eyeprocess IRT model specification and associated metadata or diagnostics needed to interpret the result.
Audit monotonicity of an item response curve
eyeprocess_irt_monotonicity_audit(theta, probability, tolerance = 1e-08)eyeprocess_irt_monotonicity_audit(theta, probability, tolerance = 1e-08)
theta |
Latent-trait value or vector of latent-trait values. |
probability |
Model-implied probability vector. |
tolerance |
Numerical or decision tolerance. |
An object of class "eye_irt_monotonicity_audit", stored as a named list, with components "monotone_non_decreasing", "n_decreases", "largest_decrease", "theta", "probability". It contains monotonicity of an item response curve and associated metadata or diagnostics needed to interpret the result.
Nominal-response category probabilities
eyeprocess_irt_nominal_probability(theta, slopes, intercepts)eyeprocess_irt_nominal_probability(theta, slopes, intercepts)
theta |
Latent-trait value or vector of latent-trait values. |
slopes |
Nominal-category slope parameters. |
intercepts |
Nominal-category intercept parameters. |
A numeric value or vector containing nominal-response category probabilities.
Audit basic plausibility of dichotomous item parameters
eyeprocess_irt_parameter_plausibility_audit( items, discrimination = c(0.2, 4), difficulty = c(-6, 6), lower_asymptote = c(0, 0.5), upper_asymptote = c(0.5, 1) )eyeprocess_irt_parameter_plausibility_audit( items, discrimination = c(0.2, 4), difficulty = c(-6, 6), lower_asymptote = c(0, 0.5), upper_asymptote = c(0.5, 1) )
items |
Item-parameter data frame or item collection. |
discrimination |
Discrimination vector or matrix. |
difficulty |
Item difficulty or location parameter. |
lower_asymptote |
Lower-asymptote parameter values. |
upper_asymptote |
Upper-asymptote parameter values. |
A data frame containing basic plausibility of dichotomous item parameters. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Computes a Bernoulli response-pattern log-likelihood standardized against its model-implied mean and variance. Extreme values are model diagnostics, not evidence of cheating, disengagement, or a psychological state.
eyeprocess_irt_person_fit_lz(observed, expected, min_probability = 1e-08)eyeprocess_irt_person_fit_lz(observed, expected, min_probability = 1e-08)
observed |
Observed responses or observed values. |
expected |
Model-expected probabilities or expected values. |
min_probability |
Lower probability bound used for numerical stabilization. |
A data frame containing standardized log-likelihood person-fit diagnostic. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compute person residual fit summaries
eyeprocess_irt_person_fit_residuals( responses, probabilities, person_ids = rownames(responses) )eyeprocess_irt_person_fit_residuals( responses, probabilities, person_ids = rownames(responses) )
responses |
Response matrix or response data. |
probabilities |
Probability matrix or vector, with dimensions appropriate to the model. |
person_ids |
Optional person identifiers. |
An object of class "eye_irt_person_fit", "data.frame", stored as a data frame, containing person residual fit summaries and associated metadata needed to interpret the result.
Draw plausible values from a discrete posterior grid
eyeprocess_irt_plausible_values(score, n = 5L, seed = 1L)eyeprocess_irt_plausible_values(score, n = 5L, seed = 1L)
score |
Score object containing posterior or uncertainty information. |
n |
Number of values, draws, or plausible values to generate. |
seed |
Random-number seed for reproducible execution. |
An R object containing plausible values from a discrete posterior grid. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Compare observed and replicated IRT discrepancy statistics
eyeprocess_irt_ppc_discrepancy( observed, replicated, statistic = c("mean_score", "score_sd", "item_means", "max_item_residual") )eyeprocess_irt_ppc_discrepancy( observed, replicated, statistic = c("mean_score", "score_sd", "item_means", "max_item_residual") )
observed |
Observed responses or observed values. |
replicated |
Replicated data or replicated statistic values. |
statistic |
Discrepancy statistic or statistic function. |
An object of class "eye_irt_ppc_discrepancy", stored as a named list, with components "statistic", "observed", "replicated", "posterior_predictive_p", "interval". It contains observed and replicated IRT discrepancy statistics and associated metadata or diagnostics needed to interpret the result.
Build a paper-ready IRT information/precision table
eyeprocess_irt_precision_evidence_table( items, theta = seq(-3, 3, by = 0.5), digits = 4L )eyeprocess_irt_precision_evidence_table( items, theta = seq(-3, 3, by = 0.5), digits = 4L )
items |
Item-parameter data frame or item collection. |
theta |
Latent-trait value or vector of latent-trait values. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready IRT information/precision table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Construct a prior-sensitivity grid for Bayesian IRT analyses
eyeprocess_irt_prior_sensitivity_grid( discrimination_scale = c(0.5, 1, 1.5), difficulty_scale = c(1, 2), guessing_mean = c(0.1, 0.2) )eyeprocess_irt_prior_sensitivity_grid( discrimination_scale = c(0.5, 1, 1.5), difficulty_scale = c(1, 2), guessing_mean = c(0.1, 0.2) )
discrimination_scale |
Prior scale for item discrimination. |
difficulty_scale |
Prior scale for item difficulty or location. |
guessing_mean |
Prior mean for the lower-asymptote or guessing parameter. |
An R object containing a prior-sensitivity grid for Bayesian IRT analyses. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarise sensitivity of an estimand across declared prior specifications
eyeprocess_irt_prior_sensitivity_summary( results, prior_id = "prior_id", estimate = "estimate" )eyeprocess_irt_prior_sensitivity_summary( results, prior_id = "prior_id", estimate = "estimate" )
results |
Results table or analysis results. |
prior_id |
Identifier for the prior specification. |
estimate |
Estimate column or numerical estimates to summarize. |
An object of class "eye_irt_prior_sensitivity", stored as a named list, with components "n_specifications", "n_finite", "median", "range", "sd", "table", "guardrail". It contains sensitivity of an estimand across declared prior specifications and associated metadata or diagnostics needed to interpret the result.
Declare prior families for Bayesian IRT engine adapters
eyeprocess_irt_prior_spec( discrimination = c("lognormal", "normal"), difficulty = "normal", guessing = c("beta", "logit-normal"), location = 0, scale = 1, guessing_shape = c(5, 17), label = "default" )eyeprocess_irt_prior_spec( discrimination = c("lognormal", "normal"), difficulty = "normal", guessing = c("beta", "logit-normal"), location = 0, scale = 1, guessing_shape = c(5, 17), label = "default" )
discrimination |
Discrimination vector or matrix. |
difficulty |
Item difficulty or location parameter. |
guessing |
Prior distribution family for the guessing parameter. |
location |
Prior location hyperparameter. |
scale |
Prior scale hyperparameter. |
guessing_shape |
Shape parameters for the guessing prior. |
label |
Human-readable label. |
An object of class "eye_irt_prior_spec", stored as a named list, with components "discrimination", "difficulty", "guessing", "location", "scale", "guessing_shape", "label". It contains declare prior families for Bayesian IRT engine adapters and associated metadata or diagnostics needed to interpret the result.
Align item parameters with process-channel summaries
eyeprocess_irt_process_alignment( item_parameters, process_profile, process_columns = NULL )eyeprocess_irt_process_alignment( item_parameters, process_profile, process_columns = NULL )
item_parameters |
Item-parameter data frame. |
process_profile |
Process-measure profile or summary. |
process_columns |
Names of process-measure columns to use. |
An object of class "eye_irt_process_alignment", stored as a named list, with components "table", "correlations", "guardrail". It contains align item parameters with process-channel summaries and associated metadata or diagnostics needed to interpret the result.
Process-aware selection penalty without mental-state inference
eyeprocess_irt_process_aware_selection_penalty( information, burden, burden_weight = 0, quality_risk = 0, quality_weight = 0 )eyeprocess_irt_process_aware_selection_penalty( information, burden, burden_weight = 0, quality_risk = 0, quality_weight = 0 )
information |
Item or test information value or vector. |
burden |
Item-level process burden or cost measure. |
burden_weight |
Weight applied to the burden penalty. |
quality_risk |
Item-level measurement-quality risk. |
quality_weight |
Weight applied to the quality-risk penalty. |
A numeric value or vector containing process-aware selection penalty without mental-state inference.
Compare psychometric DIF effect sizes with process-channel contrasts
eyeprocess_irt_process_dif_concordance( dif, process, item_id = "item_id", dif_effect = "effect", process_effect = "effect" )eyeprocess_irt_process_dif_concordance( dif, process, item_id = "item_id", dif_effect = "effect", process_effect = "effect" )
dif |
Differential-item-functioning evidence or summary. |
process |
Process-measure columns or process object. |
item_id |
Item identifier or vector of item identifiers. |
dif_effect |
DIF effect used for concordance. |
process_effect |
Process-side item effect used for concordance. |
An object of class "eye_irt_process_dif_concordance", stored as a named list, with components "n", "correlation", "table", "guardrail". It contains psychometric DIF effect sizes with process-channel contrasts and associated metadata or diagnostics needed to interpret the result.
Compute Yen-style Q3 residual correlations
eyeprocess_irt_q3(responses, probabilities, use = "pairwise.complete.obs")eyeprocess_irt_q3(responses, probabilities, use = "pairwise.complete.obs")
responses |
Response matrix or response data. |
probabilities |
Probability matrix or vector, with dimensions appropriate to the model. |
use |
Missing-data handling mode passed to the residual correlation calculation. |
An R object containing yen-style Q3 residual correlations. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Create an IRT recovery design
eyeprocess_irt_recovery_design( sample_size = c(250L, 750L), n_items = c(12L, 24L), missing_rate = c(0, 0.15), testlet_sd = c(0, 0.35), replications = 10L, seed = 20260811L )eyeprocess_irt_recovery_design( sample_size = c(250L, 750L), n_items = c(12L, 24L), missing_rate = c(0, 0.15), testlet_sd = c(0, 0.35), replications = 10L, seed = 20260811L )
sample_size |
Validation sample size or vector of sample sizes. |
n_items |
Number of items. |
missing_rate |
Proportion of responses or observations set missing. |
testlet_sd |
Standard deviation of simulated testlet effects. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed for reproducible execution. |
An object of class "eye_irt_recovery_design", "data.frame", stored as a data frame, containing an IRT recovery design and associated metadata needed to interpret the result.
Summarise recovery failure rates
eyeprocess_irt_recovery_failures(x)eyeprocess_irt_recovery_failures(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A tabular R object containing recovery failure rates; rows represent analysis units and columns contain the returned quantities.
Summarise IRT parameter recovery
eyeprocess_irt_recovery_summary(x)eyeprocess_irt_recovery_summary(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A tabular R object containing iRT parameter recovery; rows represent analysis units and columns contain the returned quantities.
Construct SBC ranks from scalar truths and posterior draws
eyeprocess_irt_sbc_ranks(truth, draws, randomize_ties = TRUE, seed = 1L)eyeprocess_irt_sbc_ranks(truth, draws, randomize_ties = TRUE, seed = 1L)
truth |
Known simulated parameter value or vector of true values. |
draws |
Posterior draws, with draws arranged by simulation case as required. |
randomize_ties |
Whether ties in SBC ranks are randomized. |
seed |
Random-number seed for reproducible execution. |
A vector or matrix containing sBC ranks from scalar truths and posterior draws, with shape determined by the supplied analysis units.
Summarise IRT SBC ranks with the package SBC diagnostics
eyeprocess_irt_sbc_summary(ranks, n_draws, bins = NULL)eyeprocess_irt_sbc_summary(ranks, n_draws, bins = NULL)
ranks |
Simulation-based-calibration rank values. |
n_draws |
Number of posterior draws underlying each rank. |
bins |
Number of bins used for rank-distribution summaries. |
An object of class "eye_irt_sbc_evidence", stored as a named list, with components "diagnostics", "ecdf_deviation", "n", "n_draws". It contains iRT SBC ranks with the package SBC diagnostics and associated metadata or diagnostics needed to interpret the result.
Score a response matrix with EAP, MAP, or ML
eyeprocess_irt_score_table( responses, items, method = c("EAP", "MAP", "ML"), person_ids = rownames(responses), ... )eyeprocess_irt_score_table( responses, items, method = c("EAP", "MAP", "ML"), person_ids = rownames(responses), ... )
responses |
Response matrix or response data. |
items |
Item-parameter data frame or item collection. |
method |
Scoring, linking, or analysis method. |
person_ids |
Optional person identifiers. |
... |
Additional arguments passed to the selected method or external engine. |
A tabular R object containing a response matrix with EAP, MAP, or ML; rows represent analysis units and columns contain the returned quantities.
Summarise score uncertainty
eyeprocess_irt_score_uncertainty(scores)eyeprocess_irt_score_uncertainty(scores)
scores |
Score object or score table. |
An object of class "eye_irt_score_uncertainty", stored as a named list, with components "n", "mean_se", "median_se", "p95_se", "marginal_reliability". It contains score uncertainty and associated metadata or diagnostics needed to interpret the result.
Summarise item-parameter drift over sessions
eyeprocess_irt_session_drift( parameters, item_id = "item_id", session = "session", parameter = "b" )eyeprocess_irt_session_drift( parameters, item_id = "item_id", session = "session", parameter = "b" )
parameters |
Item-parameter data frame or parameter estimates. |
item_id |
Item identifier or vector of item identifiers. |
session |
Session identifier or grouping variable. |
parameter |
Name of the parameter to compare. |
A tabular R object containing item-parameter drift over sessions; rows represent analysis units and columns contain the returned quantities.
Audit sparse person-item response coverage
eyeprocess_irt_sparse_design_audit( data, person, item, response = NULL, min_person_items = 3L, min_item_persons = 10L )eyeprocess_irt_sparse_design_audit( data, person, item, response = NULL, min_person_items = 3L, min_item_persons = 10L )
data |
Input data frame, matrix, or compatible analysis object. |
person |
Name of the person identifier column. |
item |
Name of the item identifier column. |
response |
Observed item response or response variable. |
min_person_items |
Minimum number of observed items required per person. |
min_item_persons |
Minimum number of observed persons required per item. |
An object of class "eye_irt_sparse_design_audit", stored as a named list, with components "n_persons", "n_items", "n_observed", "density", "person_counts", "item_counts", "sparse_persons", "sparse_items", "min_person_items", "min_item_persons". It contains sparse person-item response coverage and associated metadata or diagnostics needed to interpret the result.
Stocking-Lord characteristic-curve linking
eyeprocess_irt_stocking_lord_link( reference, focal, anchors = NULL, theta = seq(-4, 4, length.out = 81), weights = NULL, start = c(A = 1, B = 0) )eyeprocess_irt_stocking_lord_link( reference, focal, anchors = NULL, theta = seq(-4, 4, length.out = 81), weights = NULL, start = c(A = 1, B = 0) )
reference |
Reference-form or reference-group item parameters. |
focal |
Focal-form or focal-group item parameters. |
anchors |
Anchor-item identifiers. |
theta |
Latent-trait value or vector of latent-trait values. |
weights |
Optional numerical weights. |
start |
Starting values for numerical optimization. |
An object of class "eye_irt_link", stored as a named list, with components "A", "B", "method", "anchors", "objective", "convergence". It contains stocking-Lord characteristic-curve linking and associated metadata or diagnostics needed to interpret the result.
Evaluate a simple adaptive stopping rule
eyeprocess_irt_stopping_rule( n_administered, se = NA_real_, min_items = 5L, max_items = 30L, target_se = 0.3 )eyeprocess_irt_stopping_rule( n_administered, se = NA_real_, min_items = 5L, max_items = 30L, target_se = 0.3 )
n_administered |
Number of items already administered. |
se |
Standard-error values. |
min_items |
Minimum number of items required. |
max_items |
Maximum permitted test length. |
target_se |
Target conditional standard error for stopping. |
A named list with components "stop", "reason", "n_administered", "se", containing a simple adaptive stopping rule and associated metadata or diagnostics.
Compare an examinee distribution with item-bank targeting
eyeprocess_irt_targeting_gap(theta, items, breaks = seq(-4, 4, by = 0.5))eyeprocess_irt_targeting_gap(theta, items, breaks = seq(-4, 4, by = 0.5))
theta |
Latent-trait value or vector of latent-trait values. |
items |
Item-parameter data frame or item collection. |
breaks |
Break points used to summarize latent-scale targeting. |
An object of class "eye_irt_targeting_gap", stored as a named list, with components "table", "absolute_gap", "interpretation". It contains an examinee distribution with item-bank targeting and associated metadata or diagnostics needed to interpret the result.
Test characteristic curve for dichotomous item parameters
eyeprocess_irt_test_characteristic_curve(theta, items, D = 1)eyeprocess_irt_test_characteristic_curve(theta, items, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
items |
Item-parameter data frame or item collection. |
D |
Logistic scaling constant. |
An object of class "eye_irt_test_characteristic_curve", "data.frame", stored as a data frame, containing characteristic curve for dichotomous item parameters and associated metadata needed to interpret the result.
Compute a test information curve from item parameters
eyeprocess_irt_test_information(theta, items, D = 1)eyeprocess_irt_test_information(theta, items, D = 1)
theta |
Latent-trait value or vector of latent-trait values. |
items |
Item-parameter data frame or item collection. |
D |
Logistic scaling constant. |
An object of class "eye_irt_information_profile", "data.frame", stored as a data frame, containing a test information curve from item parameters and associated metadata needed to interpret the result.
Audit testlet sizes and singleton structures
eyeprocess_irt_testlet_audit(spec, min_items = 2L)eyeprocess_irt_testlet_audit(spec, min_items = 2L)
spec |
Model, validation, or analysis specification object. |
min_items |
Minimum number of items required. |
A data frame containing testlet sizes and singleton structures. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Declare a testlet structure for bifactor/two-tier IRT engines
eyeprocess_irt_testlet_spec(item_id, testlet, general_dimension = "general")eyeprocess_irt_testlet_spec(item_id, testlet, general_dimension = "general")
item_id |
Item identifier or vector of item identifiers. |
testlet |
Testlet membership identifier. |
general_dimension |
General dimension name or index. |
An object of class "eye_irt_testlet_spec", "data.frame", stored as a data frame, containing declare a testlet structure for bifactor/two-tier IRT engines and associated metadata needed to interpret the result.
Audit ordered category thresholds
eyeprocess_irt_threshold_order_audit(item_id, thresholds)eyeprocess_irt_threshold_order_audit(item_id, thresholds)
item_id |
Item identifier or vector of item identifiers. |
thresholds |
Ordered response-category thresholds. |
An object of class "eye_irt_threshold_audit", stored as a named list, with components "item_id", "thresholds", "ordered", "minimum_gap", "reversals". It contains ordered category thresholds and associated metadata or diagnostics needed to interpret the result.
Declare a response/process joint IRT specification
eyeprocess_joint_process_irt_spec( response_family = c("2pl", "rasch", "grm", "gpcm"), time_model = c("none", "lognormal", "custom"), process_channels = c("dwell", "pupil", "transitions"), person_covariates = character(), item_covariates = character(), missingness = c("ignorable", "modeled", "gated"), status = c("experimental", "reference", "gated") )eyeprocess_joint_process_irt_spec( response_family = c("2pl", "rasch", "grm", "gpcm"), time_model = c("none", "lognormal", "custom"), process_channels = c("dwell", "pupil", "transitions"), person_covariates = character(), item_covariates = character(), missingness = c("ignorable", "modeled", "gated"), status = c("experimental", "reference", "gated") )
response_family |
Response-model family. |
time_model |
Response-time model specification. |
process_channels |
Declared process-measure channels. |
person_covariates |
Person-level covariates. |
item_covariates |
Item-level covariates. |
missingness |
Missing-data handling or missingness specification. |
status |
Evidence, model, or governance status. |
An object of class "eye_joint_process_irt_spec", stored as a named list, with components "response_family", "time_model", "process_channels", "person_covariates", "item_covariates", "missingness", "status". It contains declare a response/process joint IRT specification and associated metadata or diagnostics needed to interpret the result.
Directional multidimensional 2PL information
eyeprocess_mirt_directional_information( theta, discrimination, difficulty = 0, direction = NULL, D = 1 )eyeprocess_mirt_directional_information( theta, discrimination, difficulty = 0, direction = NULL, D = 1 )
theta |
Latent-trait value or vector of latent-trait values. |
discrimination |
Discrimination vector or matrix. |
difficulty |
Item difficulty or location parameter. |
direction |
Direction vector used to project multidimensional information. |
D |
Logistic scaling constant. |
A numeric value or vector containing directional multidimensional 2PL information.
Multidimensional 2PL item information matrix
eyeprocess_mirt_information_matrix( theta, discrimination, difficulty = 0, D = 1 )eyeprocess_mirt_information_matrix( theta, discrimination, difficulty = 0, D = 1 )
theta |
Latent-trait value or vector of latent-trait values. |
discrimination |
Discrimination vector or matrix. |
difficulty |
Item difficulty or location parameter. |
D |
Logistic scaling constant. |
A numeric value or vector containing multidimensional 2PL item information matrix.
Audit multidimensional IRT loading coverage
eyeprocess_mirt_loading_audit(spec, min_items_per_dimension = 3L)eyeprocess_mirt_loading_audit(spec, min_items_per_dimension = 3L)
spec |
Model, validation, or analysis specification object. |
min_items_per_dimension |
Minimum number of items required per dimension. |
A data frame containing multidimensional IRT loading coverage. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Declare a multidimensional IRT loading structure
eyeprocess_mirt_loading_spec( items, loadings, dimension_names = colnames(loadings), simple_structure = FALSE )eyeprocess_mirt_loading_spec( items, loadings, dimension_names = colnames(loadings), simple_structure = FALSE )
items |
Item-parameter data frame or item collection. |
loadings |
Item-by-dimension loading matrix. |
dimension_names |
Optional names for latent dimensions. |
simple_structure |
Whether a simple-structure loading pattern is required. |
An object of class "eye_mirt_loading_spec", stored as a named list, with components "items", "loadings", "dimensions", "simple_structure", "violations". It contains declare a multidimensional IRT loading structure and associated metadata or diagnostics needed to interpret the result.
Construct a multichannel measurement map
eyeprocess_multichannel_measurement_map( response = "accuracy", channels = c("response_time", "dwell", "pupil", "transitions"), role = NULL )eyeprocess_multichannel_measurement_map( response = "accuracy", channels = c("response_time", "dwell", "pupil", "transitions"), role = NULL )
response |
Observed item response or response variable. |
channels |
Names or definitions of measurement channels. |
role |
Declared role of each measurement channel. |
A data frame containing a multichannel measurement map. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Declare negative-control evidence targets
eyeprocess_negative_control_evidence_plan( controls = c("permutation", "temporal_shift", "placebo_window", "known_leakage"), replications = 100L, seed = 20260811L )eyeprocess_negative_control_evidence_plan( controls = c("permutation", "temporal_shift", "placebo_window", "known_leakage"), replications = 100L, seed = 20260811L )
controls |
Negative-control methods requested by the evidence plan. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed for reproducible execution. |
An object of class "eye_negative_control_evidence_plan", stored as a named list, with components "controls", "replications", "seed", "guardrail". It contains declare negative-control evidence targets and associated metadata or diagnostics needed to interpret the result.
Build a paper-ready negative-control table
eyeprocess_negative_control_evidence_table(x, digits = 4L)eyeprocess_negative_control_evidence_table(x, digits = 4L)
x |
Object to validate, summarize, verify, or otherwise process. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready negative-control table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Prepare a sparse response/process bundle for external joint engines
eyeprocess_process_irt_data_bundle( data, person, item, response, response_time = NULL, process = character(), covariates = character() )eyeprocess_process_irt_data_bundle( data, person, item, response, response_time = NULL, process = character(), covariates = character() )
data |
Input data frame, matrix, or compatible analysis object. |
person |
Name of the person identifier column. |
item |
Name of the item identifier column. |
response |
Observed item response or response variable. |
response_time |
Response-time variable or values. |
process |
Process-measure columns or process object. |
covariates |
Optional covariate columns or covariate data. |
An object of class "eye_process_irt_data_bundle", stored as a named list, with components "data", "person", "item", "response", "response_time", "process", "covariates", "n_persons", "n_items". It contains a sparse response/process bundle for external joint engines and associated metadata or diagnostics needed to interpret the result.
Aggregate process channels by item
eyeprocess_process_item_profile(data, item, channels)eyeprocess_process_item_profile(data, item, channels)
data |
Input data frame, matrix, or compatible analysis object. |
item |
Name of the item identifier column. |
channels |
Names or definitions of measurement channels. |
A tabular R object containing aggregate process channels by item; rows represent analysis units and columns contain the returned quantities.
Summarise missingness patterns across response and process channels
eyeprocess_process_missingness_pattern(data, response, channels)eyeprocess_process_missingness_pattern(data, response, channels)
data |
Input data frame, matrix, or compatible analysis object. |
response |
Observed item response or response variable. |
channels |
Names or definitions of measurement channels. |
A data frame containing missingness patterns across response and process channels. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Aggregate process channels by person
eyeprocess_process_person_profile(data, person, channels)eyeprocess_process_person_profile(data, person, channels)
data |
Input data frame, matrix, or compatible analysis object. |
person |
Name of the person identifier column. |
channels |
Names or definitions of measurement channels. |
A tabular R object containing aggregate process channels by person; rows represent analysis units and columns contain the returned quantities.
Build a paper-ready parameter-recovery table
eyeprocess_recovery_evidence_table(x, digits = 4L)eyeprocess_recovery_evidence_table(x, digits = 4L)
x |
Object to validate, summarize, verify, or otherwise process. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready parameter-recovery table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Declare reliability evidence targets
eyeprocess_reliability_evidence_plan( metrics = c("split_half", "icc", "temporal_stability", "bland_altman"), bootstrap = 200L, seed = 20260811L )eyeprocess_reliability_evidence_plan( metrics = c("split_half", "icc", "temporal_stability", "bland_altman"), bootstrap = 200L, seed = 20260811L )
metrics |
Reliability metrics requested by the evidence plan. |
bootstrap |
Number of bootstrap replicates or bootstrap configuration. |
seed |
Random-number seed for reproducible execution. |
An object of class "eye_reliability_evidence_plan", stored as a named list, with components "metrics", "bootstrap", "seed", "guardrail". It contains declare reliability evidence targets and associated metadata or diagnostics needed to interpret the result.
Build a paper-ready reliability table
eyeprocess_reliability_evidence_table(x, digits = 4L)eyeprocess_reliability_evidence_table(x, digits = 4L)
x |
Object to validate, summarize, verify, or otherwise process. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready reliability table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarise response-time structure for joint IRT work
eyeprocess_response_time_profile(data, person, item, response_time)eyeprocess_response_time_profile(data, person, item, response_time)
data |
Input data frame, matrix, or compatible analysis object. |
person |
Name of the person identifier column. |
item |
Name of the item identifier column. |
response_time |
Response-time variable or values. |
An object of class "eye_response_time_profile", stored as a named list, with components "item", "person", "n". It contains response-time structure for joint IRT work and associated metadata or diagnostics needed to interpret the result.
Build a paper-ready SBC table
eyeprocess_sbc_evidence_table(x, digits = 4L)eyeprocess_sbc_evidence_table(x, digits = 4L)
x |
Object to validate, summarize, verify, or otherwise process. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready SBC table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Describe speed-accuracy association without causal interpretation
eyeprocess_speed_accuracy_profile(data, person, response, response_time)eyeprocess_speed_accuracy_profile(data, person, response, response_time)
data |
Input data frame, matrix, or compatible analysis object. |
person |
Name of the person identifier column. |
response |
Observed item response or response variable. |
response_time |
Response-time variable or values. |
An object of class "eye_speed_accuracy_profile", stored as a named list, with components "person", "pooled_correlation", "guardrail". It contains describe speed-accuracy association without causal interpretation and associated metadata or diagnostics needed to interpret the result.
Declare the Milestone #2 measurement-quality stress evidence plan
eyeprocess_stress_evidence_plan( missing_gaze = c(0, 0.05, 0.15, 0.3), pupil_dropout = c(0, 0.05, 0.15), calibration_offset = c(0, 0.01, 0.03, 0.06), sampling_jitter = c(0, 0.05, 0.15), aoi_label_noise = c(0, 0.02, 0.1), device_shift = c(0, 0.02, 0.05), trial_imbalance = c(0, 0.1, 0.25), seed = 20260811L )eyeprocess_stress_evidence_plan( missing_gaze = c(0, 0.05, 0.15, 0.3), pupil_dropout = c(0, 0.05, 0.15), calibration_offset = c(0, 0.01, 0.03, 0.06), sampling_jitter = c(0, 0.05, 0.15), aoi_label_noise = c(0, 0.02, 0.1), device_shift = c(0, 0.02, 0.05), trial_imbalance = c(0, 0.1, 0.25), seed = 20260811L )
missing_gaze |
Severity level for synthetic gaze missingness. |
pupil_dropout |
Severity level for synthetic pupil dropout. |
calibration_offset |
Magnitude of synthetic calibration offset. |
sampling_jitter |
Magnitude of synthetic sampling-time jitter. |
aoi_label_noise |
Rate of synthetic AOI-label corruption. |
device_shift |
Magnitude of synthetic device shift. |
trial_imbalance |
Severity of synthetic trial imbalance. |
seed |
Random-number seed for reproducible execution. |
An object of class "eye_stress_evidence_plan", stored as a named list, with components "seed". It contains declare the Milestone #2 measurement-quality stress evidence plan and associated metadata or diagnostics needed to interpret the result.
Build a paper-ready stress-test table
eyeprocess_stress_evidence_table(x, digits = 4L)eyeprocess_stress_evidence_table(x, digits = 4L)
x |
Object to validate, summarize, verify, or otherwise process. |
digits |
Number of decimal digits used for presentation. |
An R object containing a paper-ready stress-test table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarise gaps in a validation evidence atlas
eyeprocess_validation_atlas_gaps(atlas)eyeprocess_validation_atlas_gaps(atlas)
atlas |
Validation-evidence atlas object. |
A named list with components "missing_components", "unresolved_claims", "complete", containing gaps in a validation evidence atlas and associated metadata or diagnostics.
Build a machine-readable validation claim/evidence matrix
eyeprocess_validation_claim_matrix( claim_id, claim, evidence_id, evidence_type, status = "qualified", boundary = NA_character_ )eyeprocess_validation_claim_matrix( claim_id, claim, evidence_id, evidence_type, status = "qualified", boundary = NA_character_ )
claim_id |
Unique identifier for the claim. |
claim |
Text of the software-validation claim. |
evidence_id |
Identifier of evidence supporting or testing the claim. |
evidence_type |
Type or class of evidence. |
status |
Evidence, model, or governance status. |
boundary |
Explicit interpretation or scope boundary for the claim. |
A data frame containing a machine-readable validation claim/evidence matrix. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
The atlas is an organizational object. It preserves links between claims, software-validation results, figures, tables, and provenance; it does not upgrade software-validation evidence into construct-validity evidence.
eyeprocess_validation_evidence_atlas( claims, recovery = NULL, sbc = NULL, stress = NULL, reliability = NULL, negative_controls = NULL, irt = NULL, provenance = NULL, artifacts = NULL )eyeprocess_validation_evidence_atlas( claims, recovery = NULL, sbc = NULL, stress = NULL, reliability = NULL, negative_controls = NULL, irt = NULL, provenance = NULL, artifacts = NULL )
claims |
Claim-evidence mapping table. |
recovery |
Parameter-recovery evidence object or table. |
sbc |
Simulation-based-calibration evidence object or table. |
stress |
Measurement-stress evidence object or table. |
reliability |
Reliability or repeatability evidence object or table. |
negative_controls |
Negative-control evidence object or table. |
irt |
IRT-specific evidence object or table. |
provenance |
Provenance metadata or provenance object. |
artifacts |
Artifact table or file-index information. |
An object of class "eye_validation_evidence_atlas", stored as a named list, with components "claims", "components", "component_status", "coverage", "hash", "guardrail". It contains assemble a validation evidence atlas and associated metadata or diagnostics needed to interpret the result.
Grade the completeness of validation evidence
eyeprocess_validation_evidence_grade( components, required = c("design", "execution", "summary", "provenance", "hash") )eyeprocess_validation_evidence_grade( components, required = c("design", "execution", "summary", "provenance", "hash") )
components |
Named evidence components. |
required |
Required evidence components or requirements. |
An object of class "eye_validation_evidence_grade", stored as a named list, with components "grade", "required", "present", "coverage". It contains grade the completeness of validation evidence and associated metadata or diagnostics needed to interpret the result.
Create an index over frozen validation evidence artifacts
eyeprocess_validation_evidence_index(root, recursive = TRUE)eyeprocess_validation_evidence_index(root, recursive = TRUE)
root |
Root directory for evidence indexing. |
recursive |
Whether evidence files are indexed recursively. |
A data frame containing an index over frozen validation evidence artifacts. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Create an evidence manifest from files and in-memory objects
eyeprocess_validation_evidence_manifest( files = character(), objects = list(), source_commit = NA_character_, label = "eyeprocess-0.9-m2" )eyeprocess_validation_evidence_manifest( files = character(), objects = list(), source_commit = NA_character_, label = "eyeprocess-0.9-m2" )
files |
File paths to include in the evidence manifest. |
objects |
Objects to include in the evidence manifest. |
source_commit |
Source-control commit associated with the evidence. |
label |
Human-readable label. |
An object of class "eye_validation_evidence_manifest", stored as a named list, with components "label", "source_commit", "files", "objects", "generated_at". It contains an evidence manifest from files and in-memory objects and associated metadata or diagnostics needed to interpret the result.
Creates a deterministic validation plan. The plan describes software-validation scenarios and does not constitute evidence for the construct validity of any gaze, pupil, response-time, or psychometric measure.
eyeprocess_validation_plan( families = c("recovery", "sbc", "stress", "reliability", "negative_control"), sample_size = c(250L, 750L), n_items = c(12L, 24L), missing_rate = c(0, 0.15), noise_level = c("reference", "elevated"), specification = c("correct", "misspecified"), replications = 20L, seed = 20260811L, label = "eyeprocess-0.9-m2" )eyeprocess_validation_plan( families = c("recovery", "sbc", "stress", "reliability", "negative_control"), sample_size = c(250L, 750L), n_items = c(12L, 24L), missing_rate = c(0, 0.15), noise_level = c("reference", "elevated"), specification = c("correct", "misspecified"), replications = 20L, seed = 20260811L, label = "eyeprocess-0.9-m2" )
families |
Validation-evidence families to include. |
sample_size |
Validation sample size or vector of sample sizes. |
n_items |
Number of items. |
missing_rate |
Proportion of responses or observations set missing. |
noise_level |
Declared simulation noise regime. |
specification |
Whether the validation scenario is correctly specified or deliberately misspecified. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed for reproducible execution. |
label |
Human-readable label. |
An object of class "eye_validation_evidence_plan", stored as a named list, with components "families", "sample_size", "n_items", "missing_rate", "noise_level", "specification", "replications", "seed", "label". It contains declare an eyeprocess validation-evidence plan and associated metadata or diagnostics needed to interpret the result.
Evaluate readiness of a Milestone #2 validation evidence bundle
eyeprocess_validation_readiness( x, required = c("design", "recovery", "stress", "reliability", "negative_controls", "claims", "provenance") )eyeprocess_validation_readiness( x, required = c("design", "recovery", "stress", "reliability", "negative_controls", "claims", "provenance") )
x |
Object to validate, summarize, verify, or otherwise process. |
required |
Required evidence components or requirements. |
An object of class "eye_validation_readiness", stored as a named list, with components "ready", "table", "hash_valid", "source_commit". It contains readiness of a Milestone #2 validation evidence bundle and associated metadata or diagnostics needed to interpret the result.
Apply a conservative software-release evidence gate
eyeprocess_validation_release_gate( readiness, acceptance = NULL, require_hash = TRUE )eyeprocess_validation_release_gate( readiness, acceptance = NULL, require_hash = TRUE )
readiness |
Validation-readiness result. |
acceptance |
Acceptance-rule results. |
require_hash |
Whether a verified integrity hash is required. |
An object of class "eye_validation_release_gate", stored as a named list, with components "pass", "readiness", "acceptance", "hash", "interpretation". It contains a conservative software-release evidence gate and associated metadata or diagnostics needed to interpret the result.
Derive a deterministic bounded validation seed
eyeprocess_validation_seed(master_seed, index, stream = 0L)eyeprocess_validation_seed(master_seed, index, stream = 0L)
master_seed |
Master random-number seed. |
index |
Deterministic substream index. |
stream |
Named random-number stream. |
A numeric value or vector containing a deterministic bounded validation seed.
Map heterogeneous exports, register import adapters, detect formats, and combine canonical datasets.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Construct, inspect, validate, modify, and summarize canonical eye-tracking datasets.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Register and convert coordinate spaces, normalize native clocks, estimate sampling rates, and synchronize streams.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Experimental interfaces for multimodal IRT, functional pupil summaries, strategy mixtures, diffusion summaries, and missing-process sensitivity.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Persist canonical datasets, produce provenance-aware reports, and bridge gp3tools or gpbiometrics objects.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Generate declared person-item-trial features, scanpaths, transitions, entropy measures, and feature dictionaries.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Build reproducible evidence for source-format compatibility. The validation framework distinguishes declared support, synthetic-fixture testing, and empirical validation with real exports. It audits source structure, adapter detection, canonical import, field coverage, native timestamp preservation, coordinate semantics, provenance, and canonical round-trip fidelity.
validate_eye_source() validates one file or folder. Use
validation_manifest() and validate_eye_corpus() to evaluate a
versioned collection of exports from multiple vendors, devices, and software
versions. init_validation_corpus() is idempotent and preserves an existing
manifest and case files unless overwrite = TRUE is supplied. A source
can be retained temporarily in the result for creation of an anonymized
validation bundle; raw vendor data are never included in bundles by default.
anonymize_eye_dataset() replaces participant, recording, and session
identifiers, removes raw tables and source paths by default, and records the
operation in provenance. Automated anonymization cannot guarantee removal of
all study-specific free text, so exported bundles still require human review.
Depending on the function, an eye_format_validation_spec, source-file
manifest, schema-coverage table, source-preservation audit, canonical
round-trip result, eye_format_validation,
eye_corpus_validation, anonymized eye_dataset, report path, or
validation-bundle path.
read_eye_export(), validate_eye_dataset(),
supported_eye_formats(), write_eye_dataset(),
provenance_manifest()
Detect, profile, import, pair, validate, and reconstruct Gazepoint Analysis and Biometrics exports, including Gazepoint Analysis 7.x *_all_gaze.csv, *_fixations.csv, and multi-section Data_Summary_export_*.csv files.
These functions operate on the canonical relational representation used by eyeprocess. For Gazepoint Analysis 7.x exports, TIMETICK(f=...) is retained as the native clock and normalized to seconds from recording start, while media-relative TIME(...) values are retained as source fields. Fixation identifiers that restart for each media item are namespaced by stimulus. Multi-section Data Summary reports are parsed into AOI definitions, user-AOI features, raw report tables, and vendor metadata. Native fields and transformations remain represented in provenance and vendor-specific metadata.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Transform explicitly mapped data frames, CSV files, and TSV files into canonical eyeprocess datasets.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Dedicated import adapters for common and legacy eye-tracking export ecosystems.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Prepare response matrices and fit optional IRT, explanatory, response-time, DIF, shared-process, and dynamic models.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Base-R plots for traces, fixations, scanpaths, heatmaps, AOIs, pupil, biometrics, quality, timing, and models.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Declare preprocessing, filter signals, interpolate pupil gaps, detect ocular events, and retain a transformation record.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Audit data quality, flag process leakage, assess readiness, and compare preprocessing or AOI specifications.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Define, inspect, standardize, and validate the relational tables used by eyeprocess.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Simulate canonical multimodal datasets and conduct parameter-recovery or power experiments.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Reconstruct trial intervals, attach responses, register static or dynamic AOIs, and derive AOI visits.
These functions operate on the canonical relational representation used by eyeprocess. Native fields and timestamps are retained whenever possible, and transformations should be recorded in the provenance table. Optional modelling engines are used only when their packages are installed.
The returned value depends on the function. Import and transformation functions generally return an eye_dataset; audit and feature functions return data frames or enriched datasets; plotting functions return their plotted data invisibly; modelling functions return engine-specific or eyeprocess_model objects.
new_eye_dataset(), read_eye_export(), validate_eye_dataset(), provenance_manifest()
Extract facet effects from a many-facet process model
facet_effects(object, channel = c("response", "process"))facet_effects(object, channel = c("response", "process"))
object |
A fitted eyeprocess model or audit object. |
channel |
Measurement channel to inspect. |
A named list with components "random_effects", "variance_components", containing facet effects from a many-facet process model and associated metadata or diagnostics.
Classifies canonical fields after a semantic round trip. Character fields are compared exactly after NA preservation; numeric fields are compared directly and also tested for a stable affine transform.
field_fidelity_report( source, roundtrip, fields = NULL, mapping = NULL, key = NULL, tolerance = 1e-08, spec = semantic_fidelity_spec() )field_fidelity_report( source, roundtrip, fields = NULL, mapping = NULL, key = NULL, tolerance = 1e-08, spec = semantic_fidelity_spec() )
source |
Original canonical data. |
roundtrip |
Canonical data reconstructed after an interchange round trip. |
fields |
Fields to compare. Defaults to common fields. |
mapping |
Optional named character vector mapping source field names to round-trip field names. |
key |
Optional unique row-alignment fields. |
tolerance |
Numeric equality tolerance. |
spec |
A 'semantic_fidelity_spec()' object. |
An object of class 'eye_field_fidelity_report'.
Build a file hash manifest
file_hash_manifest(paths, algorithm = c("md5", "sha256"))file_hash_manifest(paths, algorithm = c("md5", "sha256"))
paths |
File paths. |
algorithm |
Hash algorithm; currently 'md5' uses base R, 'sha256' uses openssl when available. |
A data frame containing a file hash manifest. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Filter a one-dimensional eye signal robustly
filter_eye_signal( signal, width = 9L, method = c("auto", "robfilter", "runmed"), online = TRUE )filter_eye_signal( signal, width = 9L, method = c("auto", "robfilter", "runmed"), online = TRUE )
signal |
Numeric signal. |
width |
Median-filter width. |
method |
'auto', 'robfilter', or 'runmed'. |
online |
Passed to 'robfilter::med.filter()' when used. |
An 'eye_signal_filter_audit' object.
Filter pupil signal robustly
filter_pupil_signal(...)filter_pupil_signal(...)
... |
Additional arguments passed to the underlying method or helper. |
An R object containing filter pupil signal robustly. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Find process measures by channel, level, status, or text
find_process_measures( registry = process_measure_registry(), channel = NULL, level = NULL, status = NULL, query = NULL )find_process_measures( registry = process_measure_registry(), channel = NULL, level = NULL, status = NULL, query = NULL )
registry |
Registry. |
channel |
Optional channel filter. |
level |
Optional level pattern. |
status |
Optional status filter. |
query |
Optional text query. |
A tabular R object containing find process measures by channel, level, status, or text; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fingerprint_validation_case(path, algorithms = c("md5", "sha256"), include_hidden = FALSE)fingerprint_validation_case(path, algorithms = c("md5", "sha256"), include_hidden = FALSE)
path |
File or directory. |
algorithms |
Hash algorithms. Base R always supplies MD5; SHA-256 is |
|
Include hidden files. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fit a single AOI growth curve
fit_aoi_growth_curve(data, time, outcome, degree = 3L)fit_aoi_growth_curve(data, time, outcome, degree = 3L)
data |
Data frame. |
time |
Time column. |
outcome |
Numeric AOI proportion/indicator column. |
degree |
Polynomial degree. |
An object of class "eye_aoi_growth_curve", stored as a named list, with components "model", "time", "outcome", "degree", "poly", "range", "status". It contains a single AOI growth curve and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_brms_adapter(formula, data, purpose, ...)fit_brms_adapter(formula, data, purpose, ...)
formula |
Value for 'formula'. See the function description and relevant article for constraints. |
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fits the 2026 censored-normal response form item-by-item conditional on a supplied latent score. This is a useful calibration/diagnostic engine for bounded continuous process variables such as AOI proportions. It is NOT the paper's full marginal EM estimator and should therefore remain experimental.
fit_censored_normal_process_irt( response_matrix, theta, lower = 0, upper = 1, control = list(maxit = 1000) )fit_censored_normal_process_irt( response_matrix, theta, lower = 0, upper = 1, control = list(maxit = 1000) )
response_matrix |
Person x item bounded continuous matrix. |
theta |
Supplied person latent scores on the calibration scale. |
lower, upper
|
Observable bounds. |
control |
'optim()' control list. |
An object of class "eye_censored_normal_process_irt", stored as a named list, with components "coefficients", "fits", "theta", "lower", "upper", "engine", "status", "citation", "caveat". It contains conditional censored-normal calibration for bounded process measurements and associated metadata or diagnostics needed to interpret the result.
Fit a multimodal change-point IRT workflow
fit_changepoint_multimodal_irt( data, ..., gaze = "fixation_count", refit = TRUE )fit_changepoint_multimodal_irt( data, ..., gaze = "fixation_count", refit = TRUE )
data |
Input data frame or compatible tabular object. |
... |
Additional arguments passed to the selected model, engine, or method. |
gaze |
Gaze/process variable or column name. |
refit |
Whether the model is refitted after segmentation. |
An object of class "eye_changepoint_multimodal_irt", stored as a named list, with components "changepoints", "refit_requested", "status". It contains a multimodal change-point IRT workflow and associated metadata or diagnostics needed to interpret the result.
Fit a change-point RT IRT workflow
fit_changepoint_rt_irt(data, ..., refit = TRUE)fit_changepoint_rt_irt(data, ..., refit = TRUE)
data |
Input data frame or compatible tabular object. |
... |
Additional arguments passed to the selected model, engine, or method. |
refit |
Whether the model is refitted after segmentation. |
An object of class "eye_changepoint_rt_irt", stored as a named list, with components "changepoints", "refit_requested", "status". It contains a change-point RT IRT workflow and associated metadata or diagnostics needed to interpret the result.
Uses GDINA for the response layer when available and retains process features as a parallel diagnostic channel. Process/mastery associations are reported descriptively and do not redefine the Q-matrix or skill labels.
fit_cognitive_diagnosis_process( response_matrix, q_matrix, process_data = NULL, process_features = NULL, person_id = NULL, engine = c("GDINA", "external"), external_engine = NULL, ... )fit_cognitive_diagnosis_process( response_matrix, q_matrix, process_data = NULL, process_features = NULL, person_id = NULL, engine = c("GDINA", "external"), external_engine = NULL, ... )
response_matrix |
Person-by-item response matrix. |
q_matrix |
Q-matrix for cognitive-diagnosis modeling. |
process_data |
Process-data input used by the model. |
process_features |
Names of process-derived features. |
person_id |
Person or participant identifier. |
engine |
Estimation engine. |
external_engine |
Validated external fitting function. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_cognitive_diagnosis_process", stored as a named list, with components "response_model", "process_summary", "process_mastery_correlation", "q_matrix", "status". It contains cognitive-diagnosis model with process indicators and associated metadata or diagnostics needed to interpret the result.
Continuous-time IRT external-engine gate
fit_continuous_time_irt(data, external_engine = NULL, ...)fit_continuous_time_irt(data, external_engine = NULL, ...)
data |
Input data frame or compatible tabular object. |
external_engine |
Validated external fitting function. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_continuous_time_irt", stored as a named list, with components "model", "status". It contains continuous-time IRT external-engine gate and associated metadata or diagnostics needed to interpret the result.
Treats the process outcome as repeated evidence crossed by person and item, with optional contextual grouping factors. This is useful when process events themselves, not only item summaries, are the observations.
fit_crossclassified_process_irt( data, outcome, person = "participant_id", item = "item_id", context = NULL, family = c("gaussian", "binomial", "poisson", "negative_binomial"), fixed = NULL )fit_crossclassified_process_irt( data, outcome, person = "participant_id", item = "item_id", context = NULL, family = c("gaussian", "binomial", "poisson", "negative_binomial"), fixed = NULL )
data |
Input data frame or compatible tabular object. |
outcome |
Outcome variable. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
context |
Context or grouping variable. |
family |
Statistical family used by the channel or model. |
fixed |
Fixed-effects specification. |
An object of class "eye_crossclassified_process_irt", stored as a named list, with components "model", "family", "person", "item", "context", "status". It contains cross-classified process IRT reference model and associated metadata or diagnostics needed to interpret the result.
Create a gated scalable cross-classified MH-RM process IRT interface
fit_crossclassified_process_irt_mhrm(data, engine = NULL, ...)fit_crossclassified_process_irt_mhrm(data, engine = NULL, ...)
data |
Data supplied to an optional external engine. |
engine |
Optional estimator implementing the intended scalable MH-RM model. |
... |
Passed to engine. |
An object of class "eye_gated_process_model", stored as a named list, with components "id", "purpose", "required_evidence", "engine", "fit", "status", "notes", "caveat". It contains a gated scalable cross-classified MH-RM process IRT interface and associated metadata or diagnostics needed to interpret the result.
Cross-device and cross-vendor metric linking. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_device_linking(x, metric, reference_device, method = c("mixed_bland_altman", "hierarchical", "equipercentile"), device_col = "device", id_cols = c("person_id", "item_id")) apply_device_linking(x, linking_model, metric = NULL, device_col = NULL, output_col = NULL) audit_device_equivalence(x, equivalence_margin, by = c("metric", "task", "aoi")) estimate_device_specific_error(x) plot_device_agreement(x, ...) plot_device_bias_by_magnitude(x, ...) plot_device_transfer_curve(x, ...) plot_device_equivalence_intervals(x, ...) plot_cross_vendor_metric_matrix(x, ...)fit_device_linking(x, metric, reference_device, method = c("mixed_bland_altman", "hierarchical", "equipercentile"), device_col = "device", id_cols = c("person_id", "item_id")) apply_device_linking(x, linking_model, metric = NULL, device_col = NULL, output_col = NULL) audit_device_equivalence(x, equivalence_margin, by = c("metric", "task", "aoi")) estimate_device_specific_error(x) plot_device_agreement(x, ...) plot_device_bias_by_magnitude(x, ...) plot_device_transfer_curve(x, ...) plot_device_equivalence_intervals(x, ...) plot_cross_vendor_metric_matrix(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
metric |
Argument controlling 'metric'; see the function usage and returned audit metadata. |
reference_device |
Argument controlling 'reference_device'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
device_col |
Argument controlling 'device_col'; see the function usage and returned audit metadata. |
id_cols |
Argument controlling 'id_cols'; see the function usage and returned audit metadata. |
linking_model |
Argument controlling 'linking_model'; see the function usage and returned audit metadata. |
output_col |
Argument controlling 'output_col'; see the function usage and returned audit metadata. |
equivalence_margin |
Argument controlling 'equivalence_margin'; see the function usage and returned audit metadata. |
by |
Argument controlling 'by'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Fit a diffusion IRT adapter
fit_diffirt_adapter(x, model = c("D", "Q"), ...)fit_diffirt_adapter(x, model = c("D", "Q"), ...)
x |
An 'eye_dataset'. |
model |
Diffusion IRT model, '"D"' or '"Q"'. |
... |
Passed to 'diffIRT::diffIRT()'. |
An 'eyeprocess_model'.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_diffirt_engine_adapter(data, purpose, ...)fit_diffirt_engine_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Dynamic GPIRT external-engine gate
fit_dynamic_gpirt(data, external_engine = NULL, ...)fit_dynamic_gpirt(data, external_engine = NULL, ...)
data |
Input data frame or compatible tabular object. |
external_engine |
Validated external fitting function. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_dynamic_gpirt", stored as a named list, with components "model", "engine", "status". It contains dynamic GPIRT external-engine gate and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_dynamic_irtree(x, spec = dynamic_irtree_spec(), min_transitions = 10L, seed = 1L, ...)fit_dynamic_irtree(x, spec = dynamic_irtree_spec(), min_transitions = 10L, seed = 1L, ...)
x |
An 'eye_dataset' or transition/long-state data frame. |
spec |
Dynamic IRTree specification. |
min_transitions |
Minimum support per destination for baseline logits. |
seed |
Random seed for probabilistic engines. |
... |
Engine-specific arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_dynamic_irtree_stan(design, spec, seed = 1L, refresh = 0L, output_dir = NULL, ...)fit_dynamic_irtree_stan(design, spec, seed = 1L, refresh = 0L, output_dir = NULL, ...)
design |
Transition design. |
spec |
Dynamic IRTree specification. |
seed |
Random seed. |
refresh |
CmdStan refresh interval. |
output_dir |
Optional CmdStan output directory. |
... |
Additional arguments to 'CmdStanModel$sample()'. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
The built-in Cox reference conditions on an already supplied theta and is therefore an event-time *measurement diagnostic*, not a full continuous-time latent-trait estimator. A validated exact implementation can be supplied via 'external_engine'.
fit_event_time_irt( data, event_time = "event_time", event = "event", theta = "theta", person = "participant_id", item = "item_id", engine = c("cox_reference", "external"), external_engine = NULL, ... )fit_event_time_irt( data, event_time = "event_time", event = "event", theta = "theta", person = "participant_id", item = "item_id", engine = c("cox_reference", "external"), external_engine = NULL, ... )
data |
Long event/item data. |
event_time |
Time-to-event column. |
event |
Event indicator (1 event, 0 censored). |
theta |
Supplied latent-trait column for the reference engine. |
person, item
|
Person and item identifiers. |
engine |
'cox_reference' or 'external'. |
external_engine |
Function implementing a study-specific event-time IRT. |
... |
Additional arguments passed to the external engine. |
An object of class "eye_event_time_irt", stored as a named list, with components "model", "engine", "theta_conditioned", "status", "note". It contains an event-time IRT reference workflow and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_external_engine(engine, data, specification = NULL, purpose, ...)fit_external_engine(engine, data, specification = NULL, purpose, ...)
engine |
Engine name. |
data |
Engine-ready data. |
specification |
Engine-specific specification. |
purpose |
Declared scientific purpose. |
... |
Engine arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fit an eRm Rasch-family model without fallback substitution
fit_eyeprocess_erm(data, model = c("RM", "PCM"), ..., engine = "eRm")fit_eyeprocess_erm(data, model = c("RM", "PCM"), ..., engine = "eRm")
data |
Input data frame, matrix, or compatible analysis object. |
model |
Model specification passed to the selected external engine. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains an eRm Rasch-family model without fallback substitution and associated metadata or diagnostics needed to interpret the result.
Fit a G-DINA cognitive-diagnosis model without fallback substitution
fit_eyeprocess_gdina(dat, Q, model = "GDINA", ..., engine = "GDINA")fit_eyeprocess_gdina(dat, Q, model = "GDINA", ..., engine = "GDINA")
dat |
Response data supplied to the GDINA engine. |
Q |
Binary item-by-attribute Q-matrix. |
model |
Model specification passed to the selected external engine. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains a G-DINA cognitive-diagnosis model without fallback substitution and associated metadata or diagnostics needed to interpret the result.
Fit a joint response/response-time LNIRT model without fallback substitution
fit_eyeprocess_lnirt(Y, RT, quadratic = FALSE, ..., engine = "LNIRT")fit_eyeprocess_lnirt(Y, RT, quadratic = FALSE, ..., engine = "LNIRT")
Y |
Item-response matrix supplied to LNIRT. |
RT |
Response-time matrix supplied to LNIRT. |
quadratic |
Whether the LNIRT quadratic option is requested. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call", "rt_scale". It contains a joint response/response-time LNIRT model without fallback substitution and associated metadata or diagnostics needed to interpret the result.
Fit a model with mirt without substituting another estimator
fit_eyeprocess_mirt(data, model = 1, itemtype = "2PL", ..., engine = "mirt")fit_eyeprocess_mirt(data, model = 1, itemtype = "2PL", ..., engine = "mirt")
data |
Input data frame, matrix, or compatible analysis object. |
model |
Model specification passed to the selected external engine. |
itemtype |
Item type specification for mirt. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains a model with mirt without substituting another estimator and associated metadata or diagnostics needed to interpret the result.
Fit a TAM model without substituting another estimator
fit_eyeprocess_tam( resp, model = c("rasch", "2pl", "gpcm"), ..., engine = "TAM" )fit_eyeprocess_tam( resp, model = c("rasch", "2pl", "gpcm"), ..., engine = "TAM" )
resp |
Response matrix supplied to TAM. |
model |
Model specification passed to the selected external engine. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains a TAM model without substituting another estimator and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_eyetrackingr_adapter(data, purpose, ...)fit_eyetrackingr_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Spatio-temporal fixation point-process models. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_fixation_point_process(x, spatial_covariates = NULL, temporal_covariates = NULL, interaction = c("none", "self_exciting"), x_col = "x", y_col = "y", time_col = "time", grid_size = 20) fit_marked_gaze_process(x, marks = c("duration", "pupil", "saccade_amplitude"), x_col = "x", y_col = "y", time_col = "time") predict_fixation_intensity(model, new_stimulus = NULL) diagnose_gaze_point_process(model) plot_fixation_intensity(x, ...) plot_spatial_residuals(x, ...) plot_temporal_excitation_kernel(x, ...) plot_covariate_effect_surface(x, ...) plot_observed_expected_fixations(x, ...)fit_fixation_point_process(x, spatial_covariates = NULL, temporal_covariates = NULL, interaction = c("none", "self_exciting"), x_col = "x", y_col = "y", time_col = "time", grid_size = 20) fit_marked_gaze_process(x, marks = c("duration", "pupil", "saccade_amplitude"), x_col = "x", y_col = "y", time_col = "time") predict_fixation_intensity(model, new_stimulus = NULL) diagnose_gaze_point_process(model) plot_fixation_intensity(x, ...) plot_spatial_residuals(x, ...) plot_temporal_excitation_kernel(x, ...) plot_covariate_effect_surface(x, ...) plot_observed_expected_fixations(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
spatial_covariates |
Argument controlling 'spatial_covariates'; see the function usage and returned audit metadata. |
temporal_covariates |
Argument controlling 'temporal_covariates'; see the function usage and returned audit metadata. |
interaction |
Argument controlling 'interaction'; see the function usage and returned audit metadata. |
x_col |
Argument controlling 'x_col'; see the function usage and returned audit metadata. |
y_col |
Argument controlling 'y_col'; see the function usage and returned audit metadata. |
time_col |
Argument controlling 'time_col'; see the function usage and returned audit metadata. |
grid_size |
Argument controlling 'grid_size'; see the function usage and returned audit metadata. |
marks |
Argument controlling 'marks'; see the function usage and returned audit metadata. |
model |
Argument controlling 'model'; see the function usage and returned audit metadata. |
new_stimulus |
Argument controlling 'new_stimulus'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Flow-MIRT external-engine gate
fit_flow_mirt(response_matrix, external_engine = NULL, ...)fit_flow_mirt(response_matrix, external_engine = NULL, ...)
response_matrix |
Person-by-item response matrix. |
external_engine |
Validated external fitting function. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_flow_mirt", stored as a named list, with components "model", "status", "engine". It contains flow-MIRT external-engine gate and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_functional_pupil_stan(prepared, basis_matrix = NULL, seed = 1L, refresh = 0L, output_dir = NULL, ...)fit_functional_pupil_stan(prepared, basis_matrix = NULL, seed = 1L, refresh = 0L, output_dir = NULL, ...)
prepared |
Prepared functional pupil data. |
basis_matrix |
Functional basis matrix. |
seed |
Random seed. |
refresh |
CmdStan refresh interval. |
output_dir |
Optional CmdStan output directory. |
... |
Additional sampling arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fits a conventional 3PL response model with 'mirt', then aligns the fitted lower-asymptote parameter with item-level gaze/pupil/RT summaries. The alignment is descriptive and diagnostic. It must not be interpreted as a confirmatory detector of guessing, rapid responding, disengagement, or any other latent behavior without independent validation.
fit_gaze_anchored_3pl_audit( response_matrix, process_data = NULL, item = "item_id", process_features = c("ttff_ms", "dwell_ms", "pupil_bc", "pupil_peak", "rt_ms", "accuracy"), model = 1, SE = FALSE )fit_gaze_anchored_3pl_audit( response_matrix, process_data = NULL, item = "item_id", process_features = c("ttff_ms", "dwell_ms", "pupil_bc", "pupil_peak", "rt_ms", "accuracy"), model = 1, SE = FALSE )
response_matrix |
Person x item dichotomous response matrix. |
process_data |
Optional trial/person-item process table. |
item |
Item identifier in 'process_data'. |
process_features |
Candidate numeric process features. |
model |
mirt model specification. |
SE |
Request standard errors from 'mirt'. |
An 'eye_gaze_anchored_3pl_audit' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_gaze_diffusion_irt(data, spec = gaze_diffusion_spec(), seed = 1L, ...)fit_gaze_diffusion_irt(data, spec = gaze_diffusion_spec(), seed = 1L, ...)
data |
Trial-level data. |
spec |
Diffusion specification. |
seed |
Seed. |
... |
Engine arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_gaze_diffusion_stan(prepared, seed = 1L, refresh = 0L, output_dir = NULL, ...)fit_gaze_diffusion_stan(prepared, seed = 1L, refresh = 0L, output_dir = NULL, ...)
prepared |
Prepared data. |
seed |
Seed. |
refresh |
CmdStan refresh interval. |
output_dir |
Optional output directory. |
... |
Additional sampling arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fits a transparent two-part reference model: (1) whether an item response is missing and (2) the observed response, both conditional on a supplied latent trait (or a clearly labelled person-score proxy), item, and visual exposure. This is a diagnostic bridge to joint MNAR/process IRT, not a substitute for a fully joint latent missingness model.
fit_gaze_informed_missingness_irt( data, response = "response", person = "participant_id", item = "item_id", gaze_exposure = "gaze_exposure", theta = NULL, reached = NULL )fit_gaze_informed_missingness_irt( data, response = "response", person = "participant_id", item = "item_id", gaze_exposure = "gaze_exposure", theta = NULL, reached = NULL )
data |
Long person-item data. |
response |
Response column; missing values identify omissions. |
person, item
|
Person and item identifiers. |
gaze_exposure |
Non-negative visual-exposure measure. |
theta |
Optional latent-trait column. If 'NULL', a smoothed person proportion-correct logit is used as an explicit proxy. |
reached |
Optional reached/not-reached indicator. |
An 'eye_gaze_informed_missingness_irt' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_gdina_adapter(data, Q, model = "GDINA", purpose = "cognitive diagnosis", ...)fit_gdina_adapter(data, Q, model = "GDINA", purpose = "cognitive diagnosis", ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
Q |
Q-matrix with one row per item. |
model |
GDINA model specification used for eye-dataset inputs. |
purpose |
Declared scientific purpose for the external-engine contract. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
'external' is the only exact GPIRT path. 'spline_reference' fits flexible logistic spline IRFs solely as a nonparametric stress test for conventional logistic IRF shape; it is deliberately not described as a Gaussian process.
fit_gpirt( response_matrix, engine = c("spline_reference", "external"), external_engine = NULL, spline_df = 5L, ... )fit_gpirt( response_matrix, engine = c("spline_reference", "external"), external_engine = NULL, spline_df = 5L, ... )
response_matrix |
Person-by-item response matrix. |
engine |
Estimation engine. |
external_engine |
Validated external fitting function. |
spline_df |
Degrees of freedom for the spline reference model. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_gpirt", stored as a named list, with components "response_matrix", "models", "theta_proxy", "engine", "exact_gpirt", "status", "note". It contains gPIRT model-criticism interface and associated metadata or diagnostics needed to interpret the result.
Fit a registered multimodal IRT model
fit_irt_model(spec, data, ..., allow_experimental = FALSE)fit_irt_model(spec, data, ..., allow_experimental = FALSE)
spec |
IRT model or validation specification. |
data |
Input data frame or compatible tabular object. |
... |
Additional arguments passed to the selected model, engine, or method. |
allow_experimental |
Whether experimental models are permitted. |
An R object containing a registered multimodal IRT model. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Estimates screening predictions for item difficulty/discrimination from item design/process features. Predictions are not calibrated operational parameters.
fit_item_parameter_seed_model( item_data, difficulty = "irt_difficulty", discrimination = "irt_discrimination", predictors, engine = c("auto", "ranger", "lm"), seed = 2221 )fit_item_parameter_seed_model( item_data, difficulty = "irt_difficulty", discrimination = "irt_discrimination", predictors, engine = c("auto", "ranger", "lm"), seed = 2221 )
item_data |
Calibrated item-level training data. |
difficulty, discrimination
|
Target columns. |
predictors |
Design/process predictors. |
engine |
'auto', 'ranger', or 'lm'. |
seed |
Seed. |
An object of class "eye_item_parameter_seed", stored as a named list, with components "difficulty_model", "discrimination_model", "difficulty", "discrimination", "predictors", "engine", "training_data", "status", "caveat". It contains an experimental pre-pilot item-parameter seeding model and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_joint_functional_pupil_irt(x, spec = functional_pupil_irt_spec(), seed = 1L, ...)fit_joint_functional_pupil_irt(x, spec = functional_pupil_irt_spec(), seed = 1L, ...)
x |
Eye dataset or long pupil data. |
spec |
Functional pupil specification. |
seed |
Random seed. |
... |
Engine-specific arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
The reference engine fits crossed person/item submodels for accuracy, log-response-time, and a gaze process, then returns person- and item-side latent-score covariance summaries. The 'brms' engine uses multivariate formulas with shared group-level IDs so person/item random effects can be correlated across channels.
fit_joint_gaze_rt_irt( data, response = "response", rt = "rt", gaze = "fixation_count", person = "participant_id", item = "item_id", gaze_family = c("negative_binomial", "poisson"), engine = c("reference", "brms"), iter = 2000, chains = 4, cores = 1, seed = 1, ... )fit_joint_gaze_rt_irt( data, response = "response", rt = "rt", gaze = "fixation_count", person = "participant_id", item = "item_id", gaze_family = c("negative_binomial", "poisson"), engine = c("reference", "brms"), iter = 2000, chains = 4, cores = 1, seed = 1, ... )
data |
Long person-by-item data. |
response |
Binary response variable. |
rt |
Positive response-time variable. |
gaze |
Gaze process variable, usually fixation count or dwell count. |
person, item
|
Person and item identifiers. |
gaze_family |
Poisson or negative-binomial reference channel. |
engine |
'reference' or 'brms'. |
iter, chains, cores
|
Passed to brms. |
seed |
Random seed. |
... |
Additional arguments to the selected engine. |
An 'eye_joint_gaze_rt_irt' object.
Extends the 2026 graded-response/RT direction with an optional gaze/process channel. The bundled reference engine uses proportional-odds plus crossed RT and process submodels; it is explicitly experimental rather than a claim to reproduce the published SAEM estimator.
fit_joint_graded_rt_process_irt( data, response = "response", rt = "rt", process = "fixation_count", person = "participant_id", item = "item_id", engine = c("reference", "brms"), process_family = c("negative_binomial", "poisson", "gaussian"), iter = 2000, chains = 4, cores = 1, seed = 1, ... )fit_joint_graded_rt_process_irt( data, response = "response", rt = "rt", process = "fixation_count", person = "participant_id", item = "item_id", engine = c("reference", "brms"), process_family = c("negative_binomial", "poisson", "gaussian"), iter = 2000, chains = 4, cores = 1, seed = 1, ... )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
rt |
Response-time variable or column name. |
process |
Process variable or column name. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
engine |
Estimation engine. |
process_family |
Distributional family for the process channel. |
iter |
Number of estimation iterations. |
chains |
Number of Bayesian chains. |
cores |
Number of processor cores. |
seed |
Random-number seed. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_joint_graded_rt_process_irt", stored as a named list, with components "engine", "response_model", "rt_model", "process_model", "data_n", "status", "note". It contains joint graded-response, RT, and process reference model and associated metadata or diagnostics needed to interpret the result.
Create a gated KDE latent-distribution IRT interface
fit_kde_latent_distribution_irt(response_matrix, engine = NULL, ...)fit_kde_latent_distribution_irt(response_matrix, engine = NULL, ...)
response_matrix |
Response data supplied to an optional external engine. |
engine |
Optional function implementing the exact/nonparametric marginal likelihood estimator. If omitted, a gated specification is returned. |
... |
Passed to 'engine' when supplied. |
An object of class "eye_gated_process_model", stored as a named list, with components "id", "purpose", "required_evidence", "engine", "fit", "status", "notes", "caveat". It contains a gated KDE latent-distribution IRT interface and associated metadata or diagnostics needed to interpret the result.
Latent process-class IRT reference model
fit_latent_class_process_irt( data, response = "response", process_features, person = "participant_id", item = "item_id", n_classes = 2L, seed = 1 )fit_latent_class_process_irt( data, response = "response", process_features, person = "participant_id", item = "item_id", n_classes = 2L, seed = 1 )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
process_features |
Names of process-derived features. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
n_classes |
Number of latent classes. |
seed |
Random-number seed. |
An object of class "eye_latent_class_process_irt", stored as a named list, with components "response_model", "class", "centers", "data", "process_features", "status". It contains latent process-class IRT reference model and associated metadata or diagnostics needed to interpret the result.
Fit a latent-space IRT model using LSMjml
fit_latent_space_irt( response_matrix, dimensions = 2L, penalty = NULL, constraint = NULL, starts = NULL, tol = 0.001, silent = TRUE )fit_latent_space_irt( response_matrix, dimensions = 2L, penalty = NULL, constraint = NULL, starts = NULL, tol = 0.001, silent = TRUE )
response_matrix |
Person-by-item matrix with lowest score coded zero. |
dimensions |
Latent-space dimensionality. |
penalty |
Optional L2 penalty passed to 'LSMjml::LSMfit()'. |
constraint |
Optional norm constraint 'C' passed to 'LSMfit()'. |
starts |
Starting-value strategy. |
tol |
Numerical convergence tolerance. |
silent |
Whether engine messages are suppressed. |
An object of class "eye_latent_space_irt", stored as a named list, with components "model", "person_coordinates", "item_coordinates", "person_intercept", "item_intercept", "dimensions", "engine", "status". It contains a latent-space IRT model using LSMjml and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_lnirt_adapter(data, purpose, ...)fit_lnirt_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fits crossed random effects for available person, item, device, session, site, algorithm, and AOI-definition facets. For a binary response this is a generalized many-facet reference model; it is not marketed as a FACETS software replica.
fit_manyfacet_process_irt( data, response = "response", process = NULL, person = "participant_id", item = "item_id", device = NULL, session = NULL, site = NULL, algorithm = NULL, aoi_definition = NULL, process_family = c("gaussian", "poisson", "negative_binomial") )fit_manyfacet_process_irt( data, response = "response", process = NULL, person = "participant_id", item = "item_id", device = NULL, session = NULL, site = NULL, algorithm = NULL, aoi_definition = NULL, process_family = c("gaussian", "poisson", "negative_binomial") )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
process |
Process variable or column name. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
device |
Device identifier or device facet. |
session |
Session identifier or session facet. |
site |
Site identifier or site facet. |
algorithm |
Algorithm identifier or algorithm facet. |
aoi_definition |
Value supplied to 'aoi_definition'; see Details for its model-specific role. |
process_family |
Distributional family for the process channel. |
An object of class "eye_manyfacet_process_irt", stored as a named list, with components "response_model", "process_model", "facets", "process_family", "status". It contains many-facet process IRT reference model and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_mirt_adapter(data, model = 1L, purpose, ...)fit_mirt_adapter(data, model = 1L, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
model |
Value for 'model'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
This is distinct from the existing two-stage process-class reference model: the response distribution itself is calibrated using mirt's mixture density. Process features may then be compared with the fitted response classes only when a defensible class-membership extraction is available.
fit_mixture_irt_process_classes( response_matrix, n_classes = 2L, model = 1, itemtype = "2PL", SE = FALSE )fit_mixture_irt_process_classes( response_matrix, n_classes = 2L, model = 1, itemtype = "2PL", SE = FALSE )
response_matrix |
Person x item response matrix. |
n_classes |
Number of mixture classes. The currently verified internal route supports two classes ('mixture-2'). |
model |
mirt model specification, default one dimension. |
itemtype |
Item type. |
SE |
Request standard errors. |
An object of class "eye_mixture_irt_process", stored as a named list, with components "model", "coefficients", "n_classes", "itemtype", "status", "caveat". It contains a true mirt mixture-IRT response model and associated metadata or diagnostics needed to interpret the result.
Uses FactoMineR MFA when available/requested. A block-standardized PCA fallback is available as a transparent exploratory reference and is explicitly labeled.
fit_multiblock_process_map( x, blocks = NULL, id = NULL, engine = c("auto", "FactoMineR", "pca_block_scaled"), ncp = 5L )fit_multiblock_process_map( x, blocks = NULL, id = NULL, engine = c("auto", "FactoMineR", "pca_block_scaled"), ncp = 5L )
x |
A 'process_feature_blocks()' object or data frame. |
blocks |
Required if 'x' is a data frame. |
id |
Optional identifier. |
engine |
'auto', 'FactoMineR', or 'pca_block_scaled'. |
ncp |
Number of components retained where supported. |
An object of class "eye_multiblock_process_map", stored as a named list, with components "model", "person_coordinates", "variable_coordinates", "block_coordinates", "blocks", "engine", "status", "caveat". It contains a multiblock psychometric/gaze/pupil/quality structure map and associated metadata or diagnostics needed to interpret the result.
Fits the likelihood-faithful M2 reference model with CmdStanR. There is no silent fallback. Missing observations are omitted from their channel-specific likelihood under an explicit ignorable missingness assumption; the person and item structural layers remain joint across the observed channels.
fit_multimodal_m2( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", prior_profile = c("regularized", "paper_centered"), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260814L, adapt_delta = 0.95, max_treedepth = 12L, refresh = 100L, quiet_compile = TRUE )fit_multimodal_m2( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", prior_profile = c("regularized", "paper_centered"), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260814L, adapt_delta = 0.95, max_treedepth = 12L, refresh = 100L, quiet_compile = TRUE )
x |
Data frame, 'eye_multimodal_m2_simulation', or compatible eyeprocess multimodal measurement object. |
person, item, response, rt, gaze
|
Column names. |
prior_profile |
Prior profile. |
chains, parallel_chains, iter_warmup, iter_sampling
|
CmdStan sampling controls. |
seed |
Reproducibility seed. |
adapt_delta, max_treedepth, refresh
|
CmdStan controls. |
quiet_compile |
Suppress CmdStan compilation messages. |
An 'eye_multimodal_m2_fit'.
Fits the four-channel reference likelihood using CmdStanR. Pupil summaries are standardized by default for a scale-stable reference parameterization; the transformation is retained in the returned data object. No missing nuisance values are silently imputed when the corresponding nuisance column is supplied.
fit_multimodal_m3( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", pupil = "pupil", baseline = "pupil_baseline", luminance = "luminance", gaze_x = "gaze_x", gaze_y = "gaze_y", quality = "pupil_quality", time_on_task = "time_on_task", blink = "pupil_blink", interpolated = "pupil_interpolated", device = "device", session = "session", sampling_rate = "sampling_rate_hz", pupil_scale = c("z", "raw"), prior_profile = c("regularized", "paper_centered"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260815L, adapt_delta = 0.95, max_treedepth = 12L, refresh = 100L, quiet_compile = TRUE, init = 0 )fit_multimodal_m3( x, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", pupil = "pupil", baseline = "pupil_baseline", luminance = "luminance", gaze_x = "gaze_x", gaze_y = "gaze_y", quality = "pupil_quality", time_on_task = "time_on_task", blink = "pupil_blink", interpolated = "pupil_interpolated", device = "device", session = "session", sampling_rate = "sampling_rate_hz", pupil_scale = c("z", "raw"), prior_profile = c("regularized", "paper_centered"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260815L, adapt_delta = 0.95, max_treedepth = 12L, refresh = 100L, quiet_compile = TRUE, init = 0 )
x |
Data frame, M3 simulation, or compatible measurement object. |
person, item, response, rt, gaze, pupil
|
Column names. |
baseline, luminance, gaze_x, gaze_y, quality, time_on_task
|
Optional nuisance columns. |
blink, interpolated
|
Optional pupil nuisance/audit indicators. |
device, session, sampling_rate
|
Optional measurement-context audit columns. |
pupil_scale |
'"z"' or '"raw"'. |
prior_profile |
Prior profile. |
nuisance |
Named logical vector selecting the eight explicit pupil measurement nuisance terms. |
chains, parallel_chains, iter_warmup, iter_sampling
|
CmdStan sampling controls. |
seed, adapt_delta, max_treedepth, refresh, quiet_compile
|
CmdStan controls. |
init |
CmdStan initialization; default zero initializes Cholesky factors at an interior identity point. |
An 'eye_multimodal_m3_fit'.
Fits the marginalized M4 state model using CmdStanR. Sequence order is validated but never silently changed. The latent state contributes only to selected process channels in the reference implementation; the scored-response Rasch equation remains exactly state-independent.
fit_multimodal_m4( x, spec = NULL, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", pupil = "pupil", sequence = "sequence_id", order = "trial_index", baseline = "pupil_baseline", luminance = "luminance", gaze_x = "gaze_x", gaze_y = "gaze_y", quality = "pupil_quality", time_on_task = "time_on_task", blink = "pupil_blink", interpolated = "pupil_interpolated", device = "device", session = "session", sampling_rate = "sampling_rate_hz", pupil_scale = c("z", "raw"), n_states = 2L, state_channels = c("rt", "gaze", "pupil"), transition_structure = c("markov", "iid"), trait_conditioning = c("theta", "tau"), initial_trait_conditioning = TRUE, min_sequence_length = 2L, prior_profile = c("regularized", "paper_centered"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260820L, adapt_delta = 0.97, max_treedepth = 13L, refresh = 100L, quiet_compile = TRUE, init = 0 )fit_multimodal_m4( x, spec = NULL, person = "person_id", item = "item_id", response = "response", rt = "rt", gaze = "gaze", pupil = "pupil", sequence = "sequence_id", order = "trial_index", baseline = "pupil_baseline", luminance = "luminance", gaze_x = "gaze_x", gaze_y = "gaze_y", quality = "pupil_quality", time_on_task = "time_on_task", blink = "pupil_blink", interpolated = "pupil_interpolated", device = "device", session = "session", sampling_rate = "sampling_rate_hz", pupil_scale = c("z", "raw"), n_states = 2L, state_channels = c("rt", "gaze", "pupil"), transition_structure = c("markov", "iid"), trait_conditioning = c("theta", "tau"), initial_trait_conditioning = TRUE, min_sequence_length = 2L, prior_profile = c("regularized", "paper_centered"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, seed = 20260820L, adapt_delta = 0.97, max_treedepth = 13L, refresh = 100L, quiet_compile = TRUE, init = 0 )
x |
Data frame, M4 simulation, or compatible eyeprocess object. |
spec |
Optional 'multimodal_m4_spec()'. When omitted, a conservative two-state specification is created from the explicit arguments. |
person, item, response, rt, gaze, pupil
|
Column names. |
sequence, order
|
Sequence identifier and within-sequence order columns. |
baseline, luminance, gaze_x, gaze_y, quality, time_on_task, blink, interpolated
|
Optional M3 pupil nuisance columns. |
device, session, sampling_rate
|
Optional measurement-context columns. |
pupil_scale |
'"z"' or '"raw"'. |
n_states, state_channels, transition_structure, trait_conditioning
|
Used only when 'spec' is null. |
initial_trait_conditioning, min_sequence_length
|
Used only when 'spec' is null. |
prior_profile, nuisance
|
Used only when 'spec' is null. |
chains, parallel_chains, iter_warmup, iter_sampling
|
Sampling controls. |
seed, adapt_delta, max_treedepth, refresh, quiet_compile, init
|
CmdStan controls. |
An 'eye_multimodal_m4_fit' retaining data, sequence audit, specification, Stan data, sampling controls, and provenance.
Extends the gaze/RT architecture to noncognitive or other latent traits by allowing caller-defined semantic labels. The statistical engine is delegated to ‘fit_joint_gaze_rt_irt()'; interpretation remains the researcher’s job.
fit_multimodal_trait_irt( data, response, rt, gaze, person, item, trait_label = "trait", process_label = "process", ... )fit_multimodal_trait_irt( data, response, rt, gaze, person, item, trait_label = "trait", process_label = "process", ... )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
rt |
Response-time variable or column name. |
gaze |
Gaze/process variable or column name. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
trait_label |
Value supplied to 'trait_label'; see Details for its model-specific role. |
process_label |
Value supplied to 'process_label'; see Details for its model-specific role. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_joint_gaze_rt_irt", stored as a named list, with components "engine", "response_model", "rt_model", "gaze_model", "person_scores", "item_scores", "person_covariance", "item_covariance", "data_n", "gaze_family", "columns", "status", and additional components. It contains multimodal trait-model convenience wrapper and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_multinomial_transition(design, ridge = 1e-4, reference_state = NULL, control = list(maxit = 1000L, reltol = 1e-9))fit_multinomial_transition(design, ridge = 1e-4, reference_state = NULL, control = list(maxit = 1000L, reltol = 1e-9))
design |
Transition design. |
ridge |
Ridge penalty. |
reference_state |
Reference destination state. |
control |
Passed to 'optim()'. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
The bundled reference treats option selections as repeated binary outcomes with option-specific random difficulty/discrimination and optional gaze. This preserves option-level observations but does **not** reproduce the 2026 MRM/MRM-LD likelihood. Use 'engine = "external"' for a validated exact implementation of a multiple-response model with inter-option dependence.
fit_multiple_response_process_irt( data, selected = "selected", theta = "theta", person = "participant_id", item = "item_id", option = "option_id", gaze = NULL, engine = c("reference", "external"), external_engine = NULL, ... )fit_multiple_response_process_irt( data, selected = "selected", theta = "theta", person = "participant_id", item = "item_id", option = "option_id", gaze = NULL, engine = c("reference", "external"), external_engine = NULL, ... )
data |
Long person-item-option table. |
selected |
Binary option-selection indicator. |
theta |
Supplied latent-trait estimate/score. |
person, item, option
|
Identifiers. |
gaze |
Optional option-level process measure. |
engine |
'reference' or 'external'. |
external_engine |
Validated external fitter. |
... |
Arguments passed to the external engine. |
An object of class "eye_multiple_response_process_irt", stored as a named list, with components "model", "data", "gaze", "engine", "exact_multiple_response", "status", "note". It contains a multiple-response process-IRT reference model and associated metadata or diagnostics needed to interpret the result.
The bundled estimator is a transparent two-stage process-augmented nominal model: participant ability may be supplied, or a shrinkage logit accuracy proxy is estimated; option-level gaze proportions then enter a multinomial response model. This is intended for validation and exploratory distractor research, not as a replacement for a fully latent nominal-response model.
fit_nominal_gaze_irt( data, response_option = "response_option", option_gaze, person = "participant_id", item = "item_id", ability = NULL, correct_option = NULL, add_item_effects = TRUE, ... )fit_nominal_gaze_irt( data, response_option = "response_option", option_gaze, person = "participant_id", item = "item_id", ability = NULL, correct_option = NULL, add_item_effects = TRUE, ... )
data |
Input data frame or compatible tabular object. |
response_option |
Column identifying the selected response option. |
option_gaze |
Character vector naming one gaze column per response option. Names should correspond to option labels when possible. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
ability |
Optional existing ability score column. |
correct_option |
Optional scalar or column name identifying correct option. |
add_item_effects |
Whether item effects are included. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_nominal_gaze_irt", stored as a named list, with components "model", "baseline_model", "data", "option_gaze", "gaze_proportion_columns", "ability", "person", "item", "logLik_gain", "status", "note". It contains nominal/distractor IRT with option-level gaze and associated metadata or diagnostics needed to interpret the result.
Create a gated Bayesian nonignorable-missing IRT interface
fit_nonignorable_missing_irt(data, engine = NULL, ...)fit_nonignorable_missing_irt(data, engine = NULL, ...)
data |
Response/missingness data. |
engine |
Optional externally validated Bayesian estimator. |
... |
Passed to 'engine'. |
An object of class "eye_gated_process_model", stored as a named list, with components "id", "purpose", "required_evidence", "engine", "fit", "status", "notes", "caveat". It contains a gated Bayesian nonignorable-missing IRT interface and associated metadata or diagnostics needed to interpret the result.
Fits a response model among reached/answered observations and cause-specific survival models for omission and not-reached processes. The function keeps the missingness mechanisms distinct by construction.
fit_omission_survival_irt( data, response = "response", response_time = "response_time", omission_time = NULL, reached = "reached", person = "participant_id", item = "item_id", gaze_exposure = NULL, first_fixation_latency = NULL, ... )fit_omission_survival_irt( data, response = "response", response_time = "response_time", omission_time = NULL, reached = "reached", person = "participant_id", item = "item_id", gaze_exposure = NULL, first_fixation_latency = NULL, ... )
data |
Input data frame or compatible tabular object. |
response |
Response variable or response-column name. |
response_time |
Response-time variable or column name. |
omission_time |
Time associated with an omitted response. |
reached |
Indicator that the item was reached. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
gaze_exposure |
Gaze-based exposure measure. |
first_fixation_latency |
Latency to first fixation. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_omission_survival_irt", stored as a named list, with components "response_model", "omission_model", "not_reached_model", "classified_data", "state_counts", "status", "note". It contains response/RT/omission survival IRT reference model and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_openmx_adapter(model, purpose, ...)fit_openmx_adapter(model, purpose, ...)
model |
Value for 'model'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fit a user-defined OpenMx process model
fit_openmx_process_model(x, model_builder, include_features = TRUE, ...)fit_openmx_process_model(x, model_builder, include_features = TRUE, ...)
x |
An 'eye_dataset'. |
model_builder |
Function receiving model data and returning an OpenMx model. |
include_features |
Whether to merge process features. |
... |
Passed to 'OpenMx::mxRun()'. |
An 'eyeprocess_model'.
Create a gated persistence-augmented gaze-diffusion IRT interface
fit_persistence_gaze_diffusion_irt(data, engine = NULL, ...)fit_persistence_gaze_diffusion_irt(data, engine = NULL, ...)
data |
Response/RT/process data. |
engine |
Optional externally validated estimator function. |
... |
Passed to 'engine'. |
An object of class "eye_gated_process_model", stored as a named list, with components "id", "purpose", "required_evidence", "engine", "fit", "status", "notes", "caveat". It contains a gated persistence-augmented gaze-diffusion IRT interface and associated metadata or diagnostics needed to interpret the result.
Dynamic process-DIF and fairness drift. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_process_dif(x, response, process, group, item, ability = NULL) monitor_dif_drift(x, time, group, metrics, item = NULL) decompose_dif_evidence(psychometric, process = NULL, design_features = NULL) audit_fairness_transportability(x, context = "device", effect = "process_dif", item = "item_id") plot_group_icc_process_overlay(x, ...) plot_process_dif_forest(x, ...) plot_dif_drift_heatmap(x, ...) plot_fairness_transport_matrix(x, ...) plot_item_group_process_curves(x, ...)fit_process_dif(x, response, process, group, item, ability = NULL) monitor_dif_drift(x, time, group, metrics, item = NULL) decompose_dif_evidence(psychometric, process = NULL, design_features = NULL) audit_fairness_transportability(x, context = "device", effect = "process_dif", item = "item_id") plot_group_icc_process_overlay(x, ...) plot_process_dif_forest(x, ...) plot_dif_drift_heatmap(x, ...) plot_fairness_transport_matrix(x, ...) plot_item_group_process_curves(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
response |
Argument controlling 'response'; see the function usage and returned audit metadata. |
process |
Argument controlling 'process'; see the function usage and returned audit metadata. |
group |
Argument controlling 'group'; see the function usage and returned audit metadata. |
item |
Argument controlling 'item'; see the function usage and returned audit metadata. |
ability |
Argument controlling 'ability'; see the function usage and returned audit metadata. |
time |
Argument controlling 'time'; see the function usage and returned audit metadata. |
metrics |
Argument controlling 'metrics'; see the function usage and returned audit metadata. |
psychometric |
Argument controlling 'psychometric'; see the function usage and returned audit metadata. |
design_features |
Argument controlling 'design_features'; see the function usage and returned audit metadata. |
context |
Argument controlling 'context'; see the function usage and returned audit metadata. |
effect |
Argument controlling 'effect'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Process reliability and Generalizability Theory. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_process_gstudy(x, metric, facets = c("person", "item", "session", "device"), design = c("crossed", "nested")) process_variance_components(x) design_process_dstudy(gstudy, persons = NULL, items = seq(5, 50, 5), sessions = 1:5, devices = 1) audit_process_reliability(x, metrics, method = c("icc", "gtheory", "split_half", "bootstrap"), person_col = "person_id", item_col = "item_id", draws = 250) plot_variance_components(x, ...) plot_dependability_surface(x, ...) plot_reliability_by_metric(x, ...) plot_session_stability(x, ...) plot_item_sampling_reliability(x, ...)fit_process_gstudy(x, metric, facets = c("person", "item", "session", "device"), design = c("crossed", "nested")) process_variance_components(x) design_process_dstudy(gstudy, persons = NULL, items = seq(5, 50, 5), sessions = 1:5, devices = 1) audit_process_reliability(x, metrics, method = c("icc", "gtheory", "split_half", "bootstrap"), person_col = "person_id", item_col = "item_id", draws = 250) plot_variance_components(x, ...) plot_dependability_surface(x, ...) plot_reliability_by_metric(x, ...) plot_session_stability(x, ...) plot_item_sampling_reliability(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
metric |
Argument controlling 'metric'; see the function usage and returned audit metadata. |
facets |
Argument controlling 'facets'; see the function usage and returned audit metadata. |
design |
Argument controlling 'design'; see the function usage and returned audit metadata. |
gstudy |
Argument controlling 'gstudy'; see the function usage and returned audit metadata. |
persons |
Argument controlling 'persons'; see the function usage and returned audit metadata. |
items |
Argument controlling 'items'; see the function usage and returned audit metadata. |
sessions |
Argument controlling 'sessions'; see the function usage and returned audit metadata. |
devices |
Argument controlling 'devices'; see the function usage and returned audit metadata. |
metrics |
Argument controlling 'metrics'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
person_col |
Argument controlling 'person_col'; see the function usage and returned audit metadata. |
item_col |
Argument controlling 'item_col'; see the function usage and returned audit metadata. |
draws |
Argument controlling 'draws'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Fits a diagonal-Gaussian HMM to standardized process features within each sequence, then uses state occupancy as explicit process evidence in a response model. This two-stage reference engine is deliberately interpretable and should be distinguished from a fully joint HMM-IRT likelihood.
fit_process_hmm_irt( data, sequence_id = "trial_id", order = "timestamp", process_features = c("x", "y"), response = "response", person = "participant_id", item = "item_id", n_states = 3L, max_iter = 100L, tol = 1e-05, seed = 1 )fit_process_hmm_irt( data, sequence_id = "trial_id", order = "timestamp", process_features = c("x", "y"), response = "response", person = "participant_id", item = "item_id", n_states = 3L, max_iter = 100L, tol = 1e-05, seed = 1 )
data |
Input data frame or compatible tabular object. |
sequence_id |
Sequence identifier. |
order |
Within-sequence ordering variable. |
process_features |
Names of process-derived features. |
response |
Response variable or response-column name. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
n_states |
Number of latent process states. |
max_iter |
Maximum number of iterations. |
tol |
Numerical convergence tolerance. |
seed |
Random-number seed. |
An object of class "eye_process_hmm_irt", stored as a named list, with components "pi", "transition", "means", "sds", "posterior_state", "state", "row_data", "occupancy", "summary_data", "response_model", "logLik", "logLik_history", and additional components. It contains process-state HMM with an IRT response layer and associated metadata or diagnostics needed to interpret the result.
Measurement-intelligence compatibility adapters. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_process_missingness_model(x, observed, predictors, random = c("person", "item")) crossmodal_recurrence_model(x, y, outcome = NULL, channels = c("gaze_pupil", "gaze_eda", "pupil_eda"), radius = NULL, covariates = NULL)fit_process_missingness_model(x, observed, predictors, random = c("person", "item")) crossmodal_recurrence_model(x, y, outcome = NULL, channels = c("gaze_pupil", "gaze_eda", "pupil_eda"), radius = NULL, covariates = NULL)
x |
Input object or data structure appropriate for the selected analysis. |
observed |
Argument controlling 'observed'; see the function usage and returned audit metadata. |
predictors |
Argument controlling 'predictors'; see the function usage and returned audit metadata. |
random |
Argument controlling 'random'; see the function usage and returned audit metadata. |
y |
Argument controlling 'y'; see the function usage and returned audit metadata. |
outcome |
Argument controlling 'outcome'; see the function usage and returned audit metadata. |
channels |
Argument controlling 'channels'; see the function usage and returned audit metadata. |
radius |
Argument controlling 'radius'; see the function usage and returned audit metadata. |
covariates |
Argument controlling 'covariates'; see the function usage and returned audit metadata. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Conditional process reference centiles. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_process_norms(x, metric, covariates, family = c("auto", "gaussian", "lognormal")) predict_process_centiles(model, newdata, centiles = c(2.5, 10, 25, 50, 75, 90, 97.5)) score_process_deviation(model, newdata, type = c("z", "centile", "tail_probability")) audit_norm_transportability(model, new_sample) plot_process_centiles(x, ...) plot_normative_fan(x, ...) plot_person_normative_profile(x, ...) plot_item_normative_deviation(x, ...)fit_process_norms(x, metric, covariates, family = c("auto", "gaussian", "lognormal")) predict_process_centiles(model, newdata, centiles = c(2.5, 10, 25, 50, 75, 90, 97.5)) score_process_deviation(model, newdata, type = c("z", "centile", "tail_probability")) audit_norm_transportability(model, new_sample) plot_process_centiles(x, ...) plot_normative_fan(x, ...) plot_person_normative_profile(x, ...) plot_item_normative_deviation(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
metric |
Argument controlling 'metric'; see the function usage and returned audit metadata. |
covariates |
Argument controlling 'covariates'; see the function usage and returned audit metadata. |
family |
Argument controlling 'family'; see the function usage and returned audit metadata. |
model |
Argument controlling 'model'; see the function usage and returned audit metadata. |
newdata |
Argument controlling 'newdata'; see the function usage and returned audit metadata. |
centiles |
Argument controlling 'centiles'; see the function usage and returned audit metadata. |
type |
Argument controlling 'type'; see the function usage and returned audit metadata. |
new_sample |
Argument controlling 'new_sample'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Informative missingness and MNAR sensitivity. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
fit_process_observation_model(x, observed, predictors, random = c("person", "item")) fit_joint_signal_missingness(outcome, observation, method = c("selection", "shared_parameter"), x = NULL, predictors = NULL) process_pattern_mixture(x, delta = seq(-1, 1, 0.1), metric = NULL, estimand = mean) sensitivity_mnar_process(x, estimand = mean, null = 0, ...) plot_observation_probability(x, ...) plot_missingness_by_time(x, ...) plot_missingness_by_aoi(x, ...) plot_mnar_tipping_point(x, ...) plot_complete_case_sensitivity(x, ...)fit_process_observation_model(x, observed, predictors, random = c("person", "item")) fit_joint_signal_missingness(outcome, observation, method = c("selection", "shared_parameter"), x = NULL, predictors = NULL) process_pattern_mixture(x, delta = seq(-1, 1, 0.1), metric = NULL, estimand = mean) sensitivity_mnar_process(x, estimand = mean, null = 0, ...) plot_observation_probability(x, ...) plot_missingness_by_time(x, ...) plot_missingness_by_aoi(x, ...) plot_mnar_tipping_point(x, ...) plot_complete_case_sensitivity(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
observed |
Argument controlling 'observed'; see the function usage and returned audit metadata. |
predictors |
Argument controlling 'predictors'; see the function usage and returned audit metadata. |
random |
Argument controlling 'random'; see the function usage and returned audit metadata. |
outcome |
Argument controlling 'outcome'; see the function usage and returned audit metadata. |
observation |
Argument controlling 'observation'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
delta |
Argument controlling 'delta'; see the function usage and returned audit metadata. |
metric |
Argument controlling 'metric'; see the function usage and returned audit metadata. |
estimand |
Argument controlling 'estimand'; see the function usage and returned audit metadata. |
null |
Argument controlling 'null'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Fit exploratory process profiles
fit_process_profile_mixture( data, variables, k = 3L, id = "person_id", engine = c("auto", "tidyLPA", "kmeans_reference"), seed = 777 )fit_process_profile_mixture( data, variables, k = 3L, id = "person_id", engine = c("auto", "tidyLPA", "kmeans_reference"), seed = 777 )
data |
Person-level process data. |
variables |
Continuous process variables. |
k |
Number of profiles. |
id |
Optional person identifier. |
engine |
'auto', 'tidyLPA', or 'kmeans_reference'. |
seed |
Random seed. |
An object of class "eye_process_profile_mixture", stored as a named list, with components "model", "assignment", "summary", "variables", "k", "engine", "scaled_data", "status", "caveat". It contains exploratory process profiles and associated metadata or diagnostics needed to interpret the result.
Fit a process-informed Rasch tree
fit_process_rasch_tree( response_matrix, covariates, formula = NULL, maxit = 60L )fit_process_rasch_tree( response_matrix, covariates, formula = NULL, maxit = 60L )
response_matrix |
Person x item dichotomous response matrix. |
covariates |
Person-level response-process covariates. |
formula |
Optional splitting formula. If omitted, all covariates are used. |
maxit |
Maximum model iterations. |
An object of class "eye_process_rasch_tree", stored as a named list, with components "model", "covariates", "formula", "status", "caveat". It contains a process-informed Rasch tree and associated metadata or diagnostics needed to interpret the result.
Fit a luminance/fatigue/process confound model for pupil response
fit_pupil_confound_model( data, pupil = "pupil_peak", luminance = "screen_luminance", trial_order = "trial_sequence", theta = NULL, person = "person_id", item = "item_id", engine = c("auto", "mgcv", "lm") )fit_pupil_confound_model( data, pupil = "pupil_peak", luminance = "screen_luminance", trial_order = "trial_sequence", theta = NULL, person = "person_id", item = "item_id", engine = c("auto", "mgcv", "lm") )
data |
Trial-level data. |
pupil, luminance, trial_order
|
Column names. |
theta |
Optional latent-score column. |
person, item
|
Optional identifiers. |
engine |
'auto', 'mgcv', or 'lm'. |
An 'eye_pupil_confound_model' object containing raw and adjusted values.
Fits a linear superposition model of overlapping event kernels. This is a transparent reference deconvolution layer, not a universal physiological model.
fit_pupil_event_deconvolution( data, by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", events = list(stimulus = 0), tmax_ms = 930, shape = 10.1, min_samples = 20L )fit_pupil_event_deconvolution( data, by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", events = list(stimulus = 0), tmax_ms = 930, shape = 10.1, min_samples = 20L )
data |
Sample-level pupil data. |
by |
Grouping columns, typically person and trial. |
time, pupil
|
Column names. |
events |
Named list mapping event labels to scalar event times or columns. |
tmax_ms, shape
|
Kernel parameters. |
min_samples |
Minimum usable samples per group. |
An object of class "eye_pupil_deconvolution", stored as a named list, with components "fits", "effects", "fitted", "events", "tmax_ms", "shape", "by", "status", "caveat". It contains transparent event-related pupil deconvolution models and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_pupillometryr_adapter(data, purpose, ...)fit_pupillometryr_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fit an IRT response model augmented by sequence embeddings
fit_response_process_embedding_irt( data, sequences, response = "response", person = "participant_id", item = "item_id", dimensions = 5L, n = c(1L, 2L, 3L) )fit_response_process_embedding_irt( data, sequences, response = "response", person = "participant_id", item = "item_id", dimensions = 5L, n = c(1L, 2L, 3L) )
data |
Input data frame or compatible tabular object. |
sequences |
Sequence inputs. |
response |
Response variable or response-column name. |
person |
Person or participant identifier column. |
item |
Item identifier, name, or item column. |
dimensions |
Number of embedding dimensions. |
n |
Requested count or n-gram order, depending on context. |
An object of class "eye_response_process_embedding_irt", stored as a named list, with components "model", "embedding", "data", "status". It contains an IRT response model augmented by sequence embeddings and associated metadata or diagnostics needed to interpret the result.
Convenience adapter motivated by current work combining response time and item revisiting with cognitive diagnosis. The response layer is delegated to 'fit_cognitive_diagnosis_process()'; revisiting/RT/gaze remain process evidence and do not redefine attributes or the Q-matrix.
fit_revisit_process_cdm( response_matrix, q_matrix, process_data, person_id = "participant_id", revisited = "revisited", rt = "response_time", gaze = NULL, ... )fit_revisit_process_cdm( response_matrix, q_matrix, process_data, person_id = "participant_id", revisited = "revisited", rt = "response_time", gaze = NULL, ... )
response_matrix |
Person-by-item responses. |
q_matrix |
Item-by-attribute Q-matrix. |
process_data |
Long process data. |
person_id |
Person identifier in 'process_data'. |
revisited |
Revisit indicator/count column. |
rt |
Response-time column. |
gaze |
Optional gaze process columns. |
... |
Passed to 'fit_cognitive_diagnosis_process()'. |
An object of class "eye_cognitive_diagnosis_process", stored as a named list, with components "response_model", "process_summary", "process_mastery_correlation", "q_matrix", "status". It contains a revisiting-aware cognitive-diagnosis process workflow and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_seqhmm_adapter(data, purpose, ...)fit_seqhmm_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Speed-accuracy-engagement IRT convenience wrapper
fit_speed_accuracy_engagement_irt(data, ..., engine = c("reference", "brms"))fit_speed_accuracy_engagement_irt(data, ..., engine = c("reference", "brms"))
data |
Input data frame or compatible tabular object. |
... |
Additional arguments passed to the selected model, engine, or method. |
engine |
Estimation engine. |
An object of class "eye_joint_gaze_rt_irt", stored as a named list, with components "engine", "response_model", "rt_model", "gaze_model", "person_scores", "item_scores", "person_covariance", "item_covariance", "data_n", "gaze_family", "columns", "status", and additional components. It contains speed-accuracy-engagement IRT convenience wrapper and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_strategy_mixture_em(prepared, starts = prepared$spec$multiple_starts, max_iter = 300L, tolerance = 1e-7, seed = 1L)fit_strategy_mixture_em(prepared, starts = prepared$spec$multiple_starts, max_iter = 300L, tolerance = 1e-7, seed = 1L)
prepared |
Prepared strategy data. |
starts |
Number of starts. |
max_iter |
Maximum EM iterations. |
tolerance |
Relative log-likelihood tolerance. |
seed |
Seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_strategy_mixture_stan(prepared, seed = 1L, refresh = 0L, output_dir = NULL, ...)fit_strategy_mixture_stan(prepared, seed = 1L, refresh = 0L, output_dir = NULL, ...)
prepared |
Prepared strategy data. |
seed |
Seed. |
refresh |
CmdStan refresh interval. |
output_dir |
Optional output directory. |
... |
Additional CmdStan sampling arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_tam_adapter(data, purpose, ...)fit_tam_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_theory_strategy_irt(data, spec, seed = 1L, response = NULL, participant = NULL, item = NULL, ...)fit_theory_strategy_irt(data, spec, seed = 1L, response = NULL, participant = NULL, item = NULL, ...)
data |
Trial-level data. |
spec |
Strategy specification. |
seed |
Seed. |
response |
Value for 'response'. See the function description and relevant article for constraints. |
participant |
Value for 'participant'. See the function description and relevant article for constraints. |
item |
Value for 'item'. See the function description and relevant article for constraints. |
... |
Engine arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
fit_traminer_adapter(data, purpose, ...)fit_traminer_adapter(data, purpose, ...)
data |
Value for 'data'. See the function description and relevant article for constraints. |
purpose |
Value for 'purpose'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Fit one model-validation replicate
fit_validation_replicate( replicate, generator, fitter, extractor, scenario = "baseline", engine = "unspecified" )fit_validation_replicate( replicate, generator, fitter, extractor, scenario = "baseline", engine = "unspecified" )
replicate |
Replicate id. |
generator |
Function '(replicate, scenario)' returning simulated data; the simulation should expose truth via 'extract_parameter_truth()'. |
fitter |
Function accepting the simulated data (or its '$data' member). |
extractor |
Function '(fit, simulation)' returning estimates with at least 'parameter' and 'estimate'; optional 'lower'/'upper' are retained. |
scenario |
Scenario label/object passed to the generator. |
engine |
Engine label. |
A logical value or vector indicating one model-validation replicate.
Variational IRT external-engine gate
fit_variational_irt(response_matrix, external_engine = NULL, ...)fit_variational_irt(response_matrix, external_engine = NULL, ...)
response_matrix |
Person-by-item response matrix. |
external_engine |
Validated external fitting function. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_variational_irt", stored as a named list, with components "model", "status", "engine". It contains variational IRT external-engine gate and associated metadata or diagnostics needed to interpret the result.
Fit an explicit visual-context/testlet IRT model
fit_visual_context_irt( response_matrix, registry, context = NULL, itemtype = "2PL", model_dimension = "Ability", context_dimension = "VisualContextFactor", SE = FALSE )fit_visual_context_irt( response_matrix, registry, context = NULL, itemtype = "2PL", model_dimension = "Ability", context_dimension = "VisualContextFactor", SE = FALSE )
response_matrix |
Person x item response matrix. |
registry |
'visual_context_registry()' object. |
context |
Optional context identifier to model; defaults to the first shared context. |
itemtype |
mirt item type. |
model_dimension |
Name for the primary latent dimension. |
context_dimension |
Name for the context/testlet dimension. |
SE |
Request standard errors from mirt. |
An object of class "eye_visual_context_irt", stored as a named list, with components "base_model", "context_model", "comparison", "registry", "context", "positions", "itemtype", "model_string", "status", "caveat". It contains an explicit visual-context/testlet IRT model and associated metadata or diagnostics needed to interpret the result.
Distance to nearest rectangular AOI boundary
fixation_boundary_uncertainty(data, aois, x = "gaze_x", y = "gaze_y")fixation_boundary_uncertainty(data, aois, x = "gaze_x", y = "gaze_y")
data |
Gaze samples. |
aois |
Rectangular AOIs. |
x, y
|
Coordinates. |
A tabular R object containing distance to nearest rectangular AOI boundary; rows represent analysis units and columns contain the returned quantities.
Freeze IRT validation reference summaries
freeze_eyeprocess_irt_reference( recovery_summary = NULL, sbc = NULL, failures = NULL, metadata = list() )freeze_eyeprocess_irt_reference( recovery_summary = NULL, sbc = NULL, failures = NULL, metadata = list() )
recovery_summary |
Parameter-recovery summary object or table. |
sbc |
Simulation-based-calibration evidence object or table. |
failures |
Failure records or failure summary. |
metadata |
Named metadata to store with the frozen object. |
A named list with components "recovery_summary", "sbc", "failures", "metadata", "scientific_scope", containing freeze IRT validation reference summaries and associated metadata or diagnostics.
Freeze a validation atlas with a reproducibility fingerprint
freeze_eyeprocess_validation_atlas(atlas, metadata = list())freeze_eyeprocess_validation_atlas(atlas, metadata = list())
atlas |
Validation-evidence atlas object. |
metadata |
Named metadata to store with the frozen object. |
An object of class "eye_validation_atlas_freeze", stored as a named list, with components "payload", "hash", "frozen". It contains freeze a validation atlas with a reproducibility fingerprint and associated metadata or diagnostics needed to interpret the result.
Freeze a complete Milestone #2 evidence bundle
freeze_eyeprocess_validation_evidence( design, recovery = NULL, sbc = NULL, stress = NULL, reliability = NULL, negative_controls = NULL, irt = NULL, claims = NULL, provenance = NULL, source_commit = NA_character_ )freeze_eyeprocess_validation_evidence( design, recovery = NULL, sbc = NULL, stress = NULL, reliability = NULL, negative_controls = NULL, irt = NULL, claims = NULL, provenance = NULL, source_commit = NA_character_ )
design |
Validation or simulation design object. |
recovery |
Parameter-recovery evidence object or table. |
sbc |
Simulation-based-calibration evidence object or table. |
stress |
Measurement-stress evidence object or table. |
reliability |
Reliability or repeatability evidence object or table. |
negative_controls |
Negative-control evidence object or table. |
irt |
IRT-specific evidence object or table. |
claims |
Claim-evidence mapping table. |
provenance |
Provenance metadata or provenance object. |
source_commit |
Source-control commit associated with the evidence. |
An object of class "eye_validation_evidence_freeze", stored as a named list, with components "components", "presence", "source_commit", "frozen_at", "scientific_scope". It contains freeze a complete Milestone #2 evidence bundle and associated metadata or diagnostics needed to interpret the result.
Freeze software-paper evidence to an RDS with a hash
freeze_software_paper_evidence(x, path)freeze_software_paper_evidence(x, path)
x |
Evidence bundle. |
path |
Output path. |
A data frame containing freeze software-paper evidence to an RDS with a hash. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Freeze a compact validation reference for regression testing
freeze_validation_reference(x, path = NULL, digits = 8L)freeze_validation_reference(x, path = NULL, digits = 8L)
x |
Validation result. |
path |
Optional RDS path. |
digits |
Numeric rounding applied before hashing. |
An object of class "eye_validation_reference", stored as a named list, with components "summary", "failure_profile", "design_hash", "summary_hash", "created_at", "status". It contains freeze a compact validation reference for regression testing and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
functional_pupil_basis(x, df = 6L, basis = c("natural_spline", "bspline"), degree = 3L, boundary_knots = NULL, knots = NULL)functional_pupil_basis(x, df = 6L, basis = c("natural_spline", "bspline"), degree = 3L, boundary_knots = NULL, knots = NULL)
x |
Prepared functional pupil data or numeric time vector. |
df |
Degrees of freedom. |
basis |
Natural spline or B-spline. |
degree |
B-spline polynomial degree; ignored for natural splines. |
boundary_knots |
Optional boundary knots. |
knots |
Optional internal knots. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
functional_pupil_diagnostics(x)functional_pupil_diagnostics(x)
x |
Functional pupil fit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
functional_pupil_irt_spec(df = 6L, basis = c("natural_spline", "bspline"), response = "score", engine = c("two_stage_glm", "two_stage_lme4", "brms", "stan"), alignment = c("trial", "event"), event_time_column = NULL, latency_ms = 200, baseline_window = c(-200, 0), baseline_method = c("subtract", "percent", "zscore"), min_baseline_samples = 3L, drop_invalid_baseline = TRUE, time_window = NULL, pupil_column = NULL, time_column = NULL, participant_column = "participant_id", item_column = "item_id", trial_column = "trial_id", luminance_column = NULL, gaze_x_column = NULL, gaze_y_column = NULL, blink_column = NULL, interpolated_column = NULL, max_interpolated_fraction = 0.20, nuisance_by_participant = FALSE, include_response_time = TRUE, ar1 = TRUE, participant_effect = TRUE, item_effect = TRUE, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L)functional_pupil_irt_spec(df = 6L, basis = c("natural_spline", "bspline"), response = "score", engine = c("two_stage_glm", "two_stage_lme4", "brms", "stan"), alignment = c("trial", "event"), event_time_column = NULL, latency_ms = 200, baseline_window = c(-200, 0), baseline_method = c("subtract", "percent", "zscore"), min_baseline_samples = 3L, drop_invalid_baseline = TRUE, time_window = NULL, pupil_column = NULL, time_column = NULL, participant_column = "participant_id", item_column = "item_id", trial_column = "trial_id", luminance_column = NULL, gaze_x_column = NULL, gaze_y_column = NULL, blink_column = NULL, interpolated_column = NULL, max_interpolated_fraction = 0.20, nuisance_by_participant = FALSE, include_response_time = TRUE, ar1 = TRUE, participant_effect = TRUE, item_effect = TRUE, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L)
df |
Basis degrees of freedom. |
basis |
Natural spline or B-spline basis. |
response |
Response field. |
engine |
Two-stage baseline, multilevel baseline, brms bridge, or |
alignment |
Trial- or event-aligned time. |
event_time_column |
Event timestamp column required for event alignment. |
latency_ms |
Physiological latency shift applied before basis creation. |
baseline_window |
Numeric time window used for baseline correction. |
baseline_method |
Subtract, percent change, or z score. |
min_baseline_samples |
Minimum finite baseline samples per trial. |
drop_invalid_baseline |
Whether to exclude trials with invalid baselines. |
time_window |
Optional analysis time window. |
pupil_column |
Column names. |
time_column |
Column names. |
participant_column |
Column names. |
item_column |
Column names. |
trial_column |
Column names. |
luminance_column |
Optional nuisance columns. |
gaze_x_column |
Optional nuisance columns. |
gaze_y_column |
Optional nuisance columns. |
blink_column |
Optional quality columns. |
interpolated_column |
Optional quality columns. |
max_interpolated_fraction |
Maximum interpolated fraction per trial. |
nuisance_by_participant |
Whether nuisance residualization includes participant fixed effects. |
include_response_time |
Include response time in supported joint bridges. |
ar1 |
Include trial-wise AR(1) residual structure in Stan. |
participant_effect |
Include participant/item pupil effects. |
item_effect |
Include participant/item pupil effects. |
chains |
CmdStan controls. |
parallel_chains |
CmdStan controls. |
iter_warmup |
CmdStan controls. |
iter_sampling |
CmdStan controls. |
adapt_delta |
CmdStan sampler controls. |
max_treedepth |
CmdStan sampler controls. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Return the process-alignment table from a gaze-anchored 3PL audit
gaze_anchored_3pl_alignment(x)gaze_anchored_3pl_alignment(x)
x |
An 'eye_gaze_anchored_3pl_audit'. |
An R object containing return the process-alignment table from a gaze-anchored 3PL audit. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Empirical gaze data-quality profile
gaze_data_quality_profile( data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", target_x = NULL, target_y = NULL, valid = NULL, by = NULL, time_unit = c("ms", "s", "us") )gaze_data_quality_profile( data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", target_x = NULL, target_y = NULL, valid = NULL, by = NULL, time_unit = c("ms", "s", "us") )
data |
Gaze samples. |
x, y
|
Gaze coordinates. |
time |
Timestamp column. |
target_x, target_y
|
Optional known target coordinates. |
valid |
Optional validity indicator column. |
by |
Optional grouping columns. |
time_unit |
Timestamp unit. |
An object of class "eye_data_quality_profile", stored as a named list, with components "table", "coordinate_units", "caveat". It contains empirical gaze data-quality profile and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
gaze_diffusion_spec(response = "score", response_time = "response_time", participant = "participant_id", item = "item_id", drift_features = character(), boundary_features = character(), nondecision_features = character(), starting_features = character(), censor_column = NULL, contaminant = TRUE, engine = c("baseline", "stan", "ez_regression", "diffIRT", "brms"), gaze_features = NULL, chains = 4L, parallel_chains = min(4L, chains), iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.97, max_treedepth = 13L)gaze_diffusion_spec(response = "score", response_time = "response_time", participant = "participant_id", item = "item_id", drift_features = character(), boundary_features = character(), nondecision_features = character(), starting_features = character(), censor_column = NULL, contaminant = TRUE, engine = c("baseline", "stan", "ez_regression", "diffIRT", "brms"), gaze_features = NULL, chains = 4L, parallel_chains = min(4L, chains), iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.97, max_treedepth = 13L)
response |
Binary response column. |
response_time |
Response-time column in seconds. |
participant |
Participant identifier. |
item |
Item identifier. |
drift_features |
Features assigned a priori to drift rate. |
boundary_features |
Features assigned a priori to boundary separation. |
nondecision_features |
Features assigned a priori to non-decision time. |
starting_features |
Features assigned a priori to starting-point bias. |
censor_column |
Optional censoring column with 'observed', 'left', or 'right'. |
contaminant |
Whether to estimate a uniform contaminant mixture. |
engine |
Baseline approximation or Stan Wiener model. |
gaze_features |
Value for 'gaze_features'. See the function description and relevant article for constraints. |
chains |
Stan controls. |
parallel_chains |
Stan controls. |
iter_warmup |
Stan controls. |
iter_sampling |
Stan controls. |
adapt_delta |
Stan controls. |
max_treedepth |
Stan controls. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Estimate RMS successive-sample gaze imprecision
gaze_precision_rms_s2s( data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL )gaze_precision_rms_s2s( data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL )
data |
Gaze samples. |
x, y
|
Gaze-coordinate columns. |
time |
Optional timestamp column used to order samples. |
by |
Optional grouping columns. |
A logical value or vector indicating rMS successive-sample gaze imprecision.
Recurrence and cross-recurrence analysis. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
gaze_recurrence(x, representation = c("coordinates", "aoi", "velocity"), x_col = "x", y_col = "y", aoi_col = "aoi", radius = NULL) cross_recurrence(x, y, channels = c("gaze_pupil", "gaze_eda", "pupil_eda"), radius = NULL) windowed_recurrence(x, window, step) recurrence_features(x, minimum_line = 2L) plot_recurrence_matrix(x, ...) plot_windowed_recurrence(x, ...) plot_diagonal_recurrence_profile(x, ...) plot_crossmodal_recurrence(x, ...) plot_recurrence_network(x, ...)gaze_recurrence(x, representation = c("coordinates", "aoi", "velocity"), x_col = "x", y_col = "y", aoi_col = "aoi", radius = NULL) cross_recurrence(x, y, channels = c("gaze_pupil", "gaze_eda", "pupil_eda"), radius = NULL) windowed_recurrence(x, window, step) recurrence_features(x, minimum_line = 2L) plot_recurrence_matrix(x, ...) plot_windowed_recurrence(x, ...) plot_diagonal_recurrence_profile(x, ...) plot_crossmodal_recurrence(x, ...) plot_recurrence_network(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
representation |
Argument controlling 'representation'; see the function usage and returned audit metadata. |
x_col |
Argument controlling 'x_col'; see the function usage and returned audit metadata. |
y_col |
Argument controlling 'y_col'; see the function usage and returned audit metadata. |
aoi_col |
Argument controlling 'aoi_col'; see the function usage and returned audit metadata. |
radius |
Argument controlling 'radius'; see the function usage and returned audit metadata. |
y |
Argument controlling 'y'; see the function usage and returned audit metadata. |
channels |
Argument controlling 'channels'; see the function usage and returned audit metadata. |
window |
Argument controlling 'window'; see the function usage and returned audit metadata. |
step |
Argument controlling 'step'; see the function usage and returned audit metadata. |
minimum_line |
Argument controlling 'minimum_line'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Uncertainty ellipse implied by an empirical calibration-error model
gaze_uncertainty_ellipse(model, level = 0.95, center = NULL)gaze_uncertainty_ellipse(model, level = 0.95, center = NULL)
model |
Calibration error model. |
level |
Probability level. |
center |
Optional center; defaults to model mean error. |
A data frame containing uncertainty ellipse implied by an empirical calibration-error model. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Vendor-neutral preparation, validation, descriptive estimation, Cox and accelerated failure-time modelling, diagnostics, sensitivity analysis, plotting, prediction, and reporting for eye-tracking latency outcomes with explicit right censoring.
prepare_gaze_survival_data( trials, events = NULL, target_aoi = NULL, event_type = "first_fixation", participant_col = "participant_id", trial_col = "trial_id", recording_col = "recording_id", stimulus_col = "stimulus_id", condition_col = "condition", trial_start_col = NULL, trial_end_col = NULL, time_origin_col = NULL, observation_end_reason_col = NULL, event_time_col = "start_time", event_aoi_col = "aoi_id", event_trial_col = "trial_id", event_participant_col = "participant_id", event_recording_col = "recording_id", episode_type_col = "episode_type", event_observed_col = "event_observed", supplied_event_time_col = "event_time", supplied_censor_time_col = "censor_time", n_valid_samples_col = "n_valid_samples", valid_fraction_col = "valid_data_fraction", valid_observation_col = NULL, min_valid_fraction = NULL, time_origin = "trial_start", time_unit = c("seconds", "milliseconds"), source_data = NULL, preprocessing_specification = NULL, event_detector = NULL, aoi_specification = NULL, quality_rules = list() ) validate_gaze_survival_data(data, raise_on_error = TRUE) summarise_gaze_censoring(data, by = NULL) estimate_gaze_survival(data, group = NULL, conf_level = 0.95) fit_gaze_cox_model(data, formula, ties = c("breslow", "efron", "exact"), cluster = NULL) fit_gaze_mixed_cox_model( data, formula, participant_col = "participant_id", structure = NULL, ties = c("breslow", "efron", "exact") ) fit_gaze_aft_model(data, formula, distribution = NULL, ...) tidy_gaze_survival_model(model, conf_level = 0.95) check_gaze_proportional_hazards(model, transform = "km", alpha = 0.05) compare_gaze_survival_models(...) predict_gaze_survival(model, newdata, times) estimate_gaze_latency_quantiles( object, probs = c(0.25, 0.5, 0.75), group = NULL, newdata = NULL ) plot_gaze_survival_curve(data, group = NULL, ...) plot_gaze_cumulative_incidence(data, group = NULL, ...) plot_gaze_hazard(data, group = NULL, ...) plot_gaze_cox_diagnostics(model, ...) compare_gaze_survival_specifications( specifications, formula, model_families, participant_col = "participant_id", ties = "breslow" ) report_gaze_survival_model(model, conf_level = 0.95) simulate_gaze_survival_inputs( kind = c("disclosure", "verification"), seed = 20260918, n_participants = 36L, trials_per_participant = 3L ) simulate_gaze_survival_example( kind = c("disclosure", "verification"), seed = 20260918, n_participants = 36L, trials_per_participant = 3L )prepare_gaze_survival_data( trials, events = NULL, target_aoi = NULL, event_type = "first_fixation", participant_col = "participant_id", trial_col = "trial_id", recording_col = "recording_id", stimulus_col = "stimulus_id", condition_col = "condition", trial_start_col = NULL, trial_end_col = NULL, time_origin_col = NULL, observation_end_reason_col = NULL, event_time_col = "start_time", event_aoi_col = "aoi_id", event_trial_col = "trial_id", event_participant_col = "participant_id", event_recording_col = "recording_id", episode_type_col = "episode_type", event_observed_col = "event_observed", supplied_event_time_col = "event_time", supplied_censor_time_col = "censor_time", n_valid_samples_col = "n_valid_samples", valid_fraction_col = "valid_data_fraction", valid_observation_col = NULL, min_valid_fraction = NULL, time_origin = "trial_start", time_unit = c("seconds", "milliseconds"), source_data = NULL, preprocessing_specification = NULL, event_detector = NULL, aoi_specification = NULL, quality_rules = list() ) validate_gaze_survival_data(data, raise_on_error = TRUE) summarise_gaze_censoring(data, by = NULL) estimate_gaze_survival(data, group = NULL, conf_level = 0.95) fit_gaze_cox_model(data, formula, ties = c("breslow", "efron", "exact"), cluster = NULL) fit_gaze_mixed_cox_model( data, formula, participant_col = "participant_id", structure = NULL, ties = c("breslow", "efron", "exact") ) fit_gaze_aft_model(data, formula, distribution = NULL, ...) tidy_gaze_survival_model(model, conf_level = 0.95) check_gaze_proportional_hazards(model, transform = "km", alpha = 0.05) compare_gaze_survival_models(...) predict_gaze_survival(model, newdata, times) estimate_gaze_latency_quantiles( object, probs = c(0.25, 0.5, 0.75), group = NULL, newdata = NULL ) plot_gaze_survival_curve(data, group = NULL, ...) plot_gaze_cumulative_incidence(data, group = NULL, ...) plot_gaze_hazard(data, group = NULL, ...) plot_gaze_cox_diagnostics(model, ...) compare_gaze_survival_specifications( specifications, formula, model_families, participant_col = "participant_id", ties = "breslow" ) report_gaze_survival_model(model, conf_level = 0.95) simulate_gaze_survival_inputs( kind = c("disclosure", "verification"), seed = 20260918, n_participants = 36L, trials_per_participant = 3L ) simulate_gaze_survival_example( kind = c("disclosure", "verification"), seed = 20260918, n_participants = 36L, trials_per_participant = 3L )
trials |
Trial-level observation-window table. Each participant/trial key must be unique. |
events |
Optional fixation or AOI-visit event table. |
target_aoi |
Target AOI used to define the event of interest. |
event_type |
Event definition, including first fixation, first AOI entry, first evidence inspection, first revisit, first transition, or disengagement. |
participant_col, trial_col, recording_col, stimulus_col, condition_col
|
Column names defining trial identity and design fields. |
trial_start_col, trial_end_col, time_origin_col, observation_end_reason_col
|
Observation-window and time-origin columns. |
event_time_col, event_aoi_col, event_trial_col, event_participant_col, event_recording_col, episode_type_col
|
Event-table mappings. |
event_observed_col, supplied_event_time_col, supplied_censor_time_col
|
Columns used when event/censoring values are supplied directly. |
n_valid_samples_col, valid_fraction_col, valid_observation_col, min_valid_fraction
|
Quality fields and explicit minimum-validity rule. |
time_origin |
Named time origin recorded in the canonical table. |
time_unit |
Input time unit. Units are never changed silently. |
source_data, preprocessing_specification, event_detector, aoi_specification, quality_rules
|
Provenance metadata retained on derived data. |
data |
Canonical gaze-survival data. |
raise_on_error |
Whether validation errors should stop execution. |
by, group
|
Optional grouping variable for descriptive summaries or curves. |
conf_level |
Confidence level for intervals. |
formula |
Right-hand-side model formula or formula string. |
ties |
Cox tie-handling method. |
cluster |
Optional cluster column for robust Cox uncertainty. |
structure |
Required repeated-participant Cox structure: |
distribution |
Required AFT distribution: |
model, object
|
A fitted gaze-survival model or descriptive survival estimate. |
transform, alpha
|
Transformation and significance threshold used by the proportional-hazards diagnostic. |
newdata, times, probs
|
Prediction and latency-quantile inputs. |
specifications |
Named collection of explicitly prepared alternative survival datasets. |
model_families |
Explicit model families to fit across sensitivity branches. |
kind |
Synthetic example family: disclosure inspection or evidence verification. |
seed, n_participants, trials_per_participant
|
Deterministic synthetic-example controls. |
... |
Additional arguments passed to the selected backend or plotting method. |
A missing target event is not automatically censoring. A row is right-censored only
when the observation window is known and the trial is otherwise analyzable. Rows
with incomplete or unusable gaze remain explicit review states. Cox models estimate
hazard ratios; AFT models estimate multiplicative time effects. Participant-clustered
Cox uncertainty and Gaussian frailty Cox models are distinct estimators and are never
silently substituted. Information criteria from ordinary Cox partial likelihood,
coxme penalized frailty likelihood, and AFT full likelihood are not treated as
rank-comparable across likelihood bases.
Preparation returns a canonical trial-level data frame. Validation and summaries return data frames. Estimation/model helpers return survival/model objects that retain model specification and provenance. Plot helpers return their plotting objects. Synthetic helpers return deterministic raw trial/event inputs or prepared canonical data.
survival::Surv,
survival::coxph,
survival::survreg
raw <- simulate_gaze_survival_inputs(seed = 1, n_participants = 12, trials_per_participant = 3) surv <- prepare_gaze_survival_data( raw$trials, raw$events, target_aoi = "disclosure", source_data = "synthetic disclosure example", event_detector = "synthetic fixation events", aoi_specification = "disclosure rectangle" ) summarise_gaze_censoring(surv, by = "condition") validate_gaze_survival_data(surv, raise_on_error = FALSE) if (requireNamespace("survival", quietly = TRUE)) { km <- estimate_gaze_survival(surv, group = "condition") cox <- fit_gaze_mixed_cox_model( surv, ~ condition, participant_col = "participant_id", structure = "cluster_robust" ) tidy_gaze_survival_model(cox) }raw <- simulate_gaze_survival_inputs(seed = 1, n_participants = 12, trials_per_participant = 3) surv <- prepare_gaze_survival_data( raw$trials, raw$events, target_aoi = "disclosure", source_data = "synthetic disclosure example", event_detector = "synthetic fixation events", aoi_specification = "disclosure rectangle" ) summarise_gaze_censoring(surv, by = "condition") validate_gaze_survival_data(surv, raise_on_error = FALSE) if (requireNamespace("survival", quietly = TRUE)) { km <- estimate_gaze_survival(surv, group = "condition") cox <- fit_gaze_mixed_cox_model( surv, ~ condition, participant_col = "participant_id", structure = "cluster_robust" ) tidy_gaze_survival_model(cox) }
Build person-by-item-by-trial analysis tables
gazepoint_analysis_tables(x)gazepoint_analysis_tables(x)
x |
A processed 'eye_dataset' with workflow features. |
A named list containing trial, AOI, fixation, pupil, biometric, process, and feature-dictionary tables.
This function does not manufacture responses. When no responses have been supplied, it returns a complete response template and marks the outcome as process-ready but response-pending.
gazepoint_irt_tables(x, process_table = NULL)gazepoint_irt_tables(x, process_table = NULL)
x |
A processed 'eye_dataset'. |
process_table |
Optional process table from 'gazepoint_analysis_tables()'. |
A list containing a response template, process covariates, long IRT table, optional matrices, and a readiness assessment.
Creates a declarative specification for the end-to-end Gazepoint workflow. The defaults preserve vendor fixations, interpolate only short pupil gaps, apply a small median pupil filter, and avoid automatic pupil baseline correction when no pre-stimulus baseline is available.
gazepoint_workflow_spec( expected_sampling_rate = 60, sampling_tolerance_hz = 5, minimum_valid_gaze = 0.8, minimum_valid_pupil = 0.7, pupil_interpolation = "linear", pupil_max_gap_ms = 150, pupil_filter = "median", pupil_window = 5L, pupil_baseline = "none", pupil_baseline_window = c(0, 0.5), detect_blinks = TRUE, biometric_channels = c("eda", "skin_conductance_level", "skin_conductance_response", "heart_rate", "interbeat_interval", "engagement_dial"), create_plots = TRUE, create_html_report = TRUE, retain_raw = TRUE )gazepoint_workflow_spec( expected_sampling_rate = 60, sampling_tolerance_hz = 5, minimum_valid_gaze = 0.8, minimum_valid_pupil = 0.7, pupil_interpolation = "linear", pupil_max_gap_ms = 150, pupil_filter = "median", pupil_window = 5L, pupil_baseline = "none", pupil_baseline_window = c(0, 0.5), detect_blinks = TRUE, biometric_channels = c("eda", "skin_conductance_level", "skin_conductance_response", "heart_rate", "interbeat_interval", "engagement_dial"), create_plots = TRUE, create_html_report = TRUE, retain_raw = TRUE )
expected_sampling_rate |
Expected gaze sampling rate in hertz. |
sampling_tolerance_hz |
Allowed absolute sampling-rate deviation. |
minimum_valid_gaze |
Minimum acceptable valid-gaze fraction. |
minimum_valid_pupil |
Minimum acceptable valid-pupil fraction. |
pupil_interpolation |
Pupil interpolation method. |
pupil_max_gap_ms |
Maximum pupil gap eligible for interpolation. |
pupil_filter |
Pupil smoothing method. |
pupil_window |
Smoothing window in samples. |
pupil_baseline |
Baseline correction method. The default is '"none"'. |
pupil_baseline_window |
Baseline window relative to media/trial onset. |
detect_blinks |
Whether to derive blink episodes from missing pupil data. |
biometric_channels |
Channels to retain in workflow plots and tables. |
create_plots |
Whether to create the complete plot suite. |
create_html_report |
Whether to render an HTML copy of the report when 'rmarkdown' and Pandoc are available. |
retain_raw |
Whether imported native exports are retained in the object. |
An 'eye_gazepoint_workflow_spec' object.
Useful before many-facet IRT: quantifies how much variance comes from person, item, device, session, algorithm, and residual sources.
generalizability_process_study(data, outcome, facets, REML = TRUE)generalizability_process_study(data, outcome, facets, REML = TRUE)
data |
Input data frame or compatible tabular object. |
outcome |
Outcome variable. |
facets |
Facet variables included in the analysis. |
REML |
Whether restricted maximum likelihood is used. |
An object of class "eye_process_g_study", stored as a named list, with components "model", "variance_components", "facets", "outcome". It contains generalizability-style variance decomposition for a process measure and associated metadata or diagnostics needed to interpret the result.
Retrieve a registered multimodal IRT model
get_irt_model(id)get_irt_model(id)
id |
Stable identifier. |
An R object containing retrieve a registered multimodal IRT model. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
This is intentionally conservative: failure of any required criterion caps the evidence grade. It does not convert exploratory evidence into a claim of substantive validity.
grade_model_evidence( recovery, spec = irt_validation_spec("unspecified"), external_validation = NULL, sbc = NULL, ppc = NULL, semantic_roundtrip = NULL )grade_model_evidence( recovery, spec = irt_validation_spec("unspecified"), external_validation = NULL, sbc = NULL, ppc = NULL, semantic_roundtrip = NULL )
recovery |
Value supplied to 'recovery'; see Details for its model-specific role. |
spec |
IRT model or validation specification. |
external_validation |
Value supplied to 'external_validation'; see Details for its model-specific role. |
sbc |
Value supplied to 'sbc'; see Details for its model-specific role. |
ppc |
Value supplied to 'ppc'; see Details for its model-specific role. |
semantic_roundtrip |
Value supplied to 'semantic_roundtrip'; see Details for its model-specific role. |
An object of class "eye_irt_evidence_grade", stored as a named list, with components "model_id", "grade", "checks", "recovery", "contract", "warning". It contains grade model evidence against an explicit validation contract and associated metadata or diagnostics needed to interpret the result.
Evaluate a model with grouped cross-validation
grouped_cv( data, formula, family = stats::binomial(), group = "participant_id", v = 5L, metric = c("log_loss", "brier", "accuracy"), seed = 1L )grouped_cv( data, formula, family = stats::binomial(), group = "participant_id", v = 5L, metric = c("log_loss", "brier", "accuracy"), seed = 1L )
data |
Data frame. |
formula |
Model formula. |
family |
GLM family. |
group |
Grouping columns. |
v |
Number of folds. |
metric |
Metric: log loss, Brier score, or accuracy. |
seed |
Random seed. |
An 'eye_grouped_cv' object.
Create grouped cross-validation folds
grouped_folds(data, group = c("participant_id"), v = 5L, seed = 1L)grouped_folds(data, group = c("participant_id"), v = 5L, seed = 1L)
data |
Data frame. |
group |
Grouping columns that must not cross folds. |
v |
Number of folds. |
seed |
Random seed. |
An 'eye_grouped_folds' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
import_benchmark_study(study = eyeprocess_benchmark_study())import_benchmark_study(study = eyeprocess_benchmark_study())
study |
Benchmark object or path. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Import Eye-Tracking-BIDS physiological recordings
import_eye_bids(path, validate = TRUE)import_eye_bids(path, validate = TRUE)
path |
BIDS dataset root. |
validate |
Whether to validate the resulting canonical dataset. |
An 'eye_dataset'.
Extract incremental process validity
incremental_process_validity(x)incremental_process_validity(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing incremental process validity. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
init_vendor_corpus(path, overwrite = FALSE)init_vendor_corpus(path, overwrite = FALSE)
path |
Corpus directory. |
overwrite |
Reinitialize an existing empty corpus. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Inject AOI label noise
inject_aoi_label_noise(data, aoi = "aoi", proportion, seed = 1L)inject_aoi_label_noise(data, aoi = "aoi", proportion, seed = 1L)
data |
Data. |
aoi |
AOI label column. |
proportion |
Proportion reassigned to another observed label. |
seed |
Seed. |
An R object containing inject AOI label noise. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject additive gaze calibration offset in coordinate units
inject_calibration_offset( data, x = "gaze_x", y = "gaze_y", offset_x = 0, offset_y = 0 )inject_calibration_offset( data, x = "gaze_x", y = "gaze_y", offset_x = 0, offset_y = 0 )
data |
Data. |
x, y
|
Gaze coordinate columns. |
offset_x, offset_y
|
Additive offsets in the same units as x/y. |
An R object containing inject additive gaze calibration offset in coordinate units. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject an additive device/site shift in a numeric feature
inject_device_shift(data, column, shift, rows = NULL)inject_device_shift(data, column, shift, rows = NULL)
data |
Data. |
column |
Numeric column. |
shift |
Additive shift. |
rows |
Optional logical/index rows affected. |
An R object containing inject an additive device/site shift in a numeric feature. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject generic missingness into selected columns
inject_eye_missingness(data, columns, proportion, seed = 1L)inject_eye_missingness(data, columns, proportion, seed = 1L)
data |
Data. |
columns |
Columns. |
proportion |
Missingness proportion. |
seed |
Seed. |
An R object containing inject generic missingness into selected columns. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject pupil dropout
inject_pupil_dropout(data, pupil = "pupil", proportion, seed = 1L)inject_pupil_dropout(data, pupil = "pupil", proportion, seed = 1L)
data |
Data. |
pupil |
Pupil column. |
proportion |
Dropout proportion. |
seed |
Seed. |
An R object containing inject pupil dropout. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject timestamp jitter
inject_sampling_jitter(data, time = "timestamp_ms", sd, seed = 1L)inject_sampling_jitter(data, time = "timestamp_ms", sd, seed = 1L)
data |
Data. |
time |
Timestamp column. |
sd |
Jitter standard deviation in timestamp units. |
seed |
Seed. |
An R object containing inject timestamp jitter. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Inject trial/row imbalance by dropping observations
inject_trial_imbalance(data, proportion, seed = 1L)inject_trial_imbalance(data, proportion, seed = 1L)
data |
Data. |
proportion |
Proportion dropped. |
seed |
Seed. |
A tabular R object containing inject trial/row imbalance by dropping observations; rows represent analysis units and columns contain the returned quantities.
Compositional AOI channel
irt_compositional_channel( parts, family = c("logratio_gaussian", "dirichlet"), latent = "process", options = list() )irt_compositional_channel( parts, family = c("logratio_gaussian", "dirichlet"), latent = "process", options = list() )
parts |
Compositional parts. |
family |
Statistical family used by the channel or model. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing compositional AOI channel and associated metadata or diagnostics.
Continuous/bounded process channel for multimodal IRT
irt_continuous_channel( family = c("censored_normal", "beta", "gaussian"), value = "process_value", lower = 0, upper = 1, latent = "process", options = list() )irt_continuous_channel( family = c("censored_normal", "beta", "gaussian"), value = "process_value", lower = 0, upper = 1, latent = "process", options = list() )
family |
Continuous response family. |
value |
Variable name. |
lower, upper
|
Bounds where applicable. |
latent |
Latent dimension. |
options |
Additional channel metadata. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing continuous/bounded process channel for multimodal IRT and associated metadata or diagnostics.
Count-valued process channel for multimodal IRT
irt_count_channel( family = c("negative_binomial", "poisson"), value = "fixation_count", latent = "engagement", options = list() )irt_count_channel( family = c("negative_binomial", "poisson"), value = "fixation_count", latent = "engagement", options = list() )
family |
Statistical family used by the channel or model. |
value |
Process-value column or values. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing count-valued process channel for multimodal IRT and associated metadata or diagnostics.
Functional trajectory channel
irt_functional_channel( value = "pupil", time = "time", family = c("basis_gaussian", "functional_factor"), latent = "process", options = list() )irt_functional_channel( value = "pupil", time = "time", family = c("basis_gaussian", "functional_factor"), latent = "process", options = list() )
value |
Process-value column or values. |
time |
Time values. |
family |
Statistical family used by the channel or model. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing functional trajectory channel and associated metadata or diagnostics.
Define a multimodal IRT model specification
irt_model_spec( id, latent, channels, status = c("experimental", "reference", "gated"), fit_fun = NULL, simulate_fun = NULL, validate_fun = NULL, citation = character(), description = NULL, requirements = character(), metadata = list() )irt_model_spec( id, latent, channels, status = c("experimental", "reference", "gated"), fit_fun = NULL, simulate_fun = NULL, validate_fun = NULL, citation = character(), description = NULL, requirements = character(), metadata = list() )
id |
Stable model identifier. |
latent |
Named or unnamed latent dimensions. |
channels |
Named list of channel objects. |
status |
One of reference, experimental, or gated. |
fit_fun |
Optional fitting function. |
simulate_fun |
Optional simulator. |
validate_fun |
Optional model-specific validation function. |
citation |
Character vector of citations/DOIs. |
description |
Human-readable model description. |
requirements |
Optional packages/engines. |
metadata |
Additional metadata. |
An 'eye_irt_model_spec'.
Nominal response/process channel
irt_nominal_channel( choice = "response_option", categories = NULL, latent = "ability", options = list() )irt_nominal_channel( choice = "response_option", categories = NULL, latent = "ability", options = list() )
choice |
Observed nominal choice. |
categories |
Nominal response categories. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing nominal response/process channel and associated metadata or diagnostics.
Binary/ordinal response channel for multimodal IRT
irt_response_channel( family = c("2pl", "rasch", "graded", "partial_credit"), response = "response", latent = "ability", options = list() )irt_response_channel( family = c("2pl", "rasch", "graded", "partial_credit"), response = "response", latent = "ability", options = list() )
family |
Statistical family used by the channel or model. |
response |
Response variable or response-column name. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing binary/ordinal response channel for multimodal IRT and associated metadata or diagnostics.
Response-time channel for multimodal IRT
irt_rt_channel( family = c("lognormal", "gaussian_log", "shifted_lognormal"), rt = "rt", latent = "speed", options = list() )irt_rt_channel( family = c("lognormal", "gaussian_log", "shifted_lognormal"), rt = "rt", latent = "speed", options = list() )
family |
Statistical family used by the channel or model. |
rt |
Response-time variable or column name. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing response-time channel for multimodal IRT and associated metadata or diagnostics.
Sequence/process-state channel
irt_sequence_channel( sequence = "sequence", family = c("ngram", "hmm", "embedding"), latent = "strategy", options = list() )irt_sequence_channel( sequence = "sequence", family = c("ngram", "hmm", "embedding"), latent = "strategy", options = list() )
sequence |
Sequence input. |
family |
Statistical family used by the channel or model. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing sequence/process-state channel and associated metadata or diagnostics.
Survival/event-time channel for multimodal IRT
irt_survival_channel( family = c("cox", "weibull", "exponential"), time = "time", event = "event", latent = NULL, options = list() )irt_survival_channel( family = c("cox", "weibull", "exponential"), time = "time", event = "event", latent = NULL, options = list() )
family |
Statistical family used by the channel or model. |
time |
Time values. |
event |
Event indicator or event column. |
latent |
Latent variable or latent-variable labels. |
options |
Additional channel/model options. |
A named list with components "type", "family", "role", "link", "variables", "latent", "options", containing survival/event-time channel for multimodal IRT and associated metadata or diagnostics.
Creates an evidence contract rather than executing a particular estimator. The object can be passed to stress-test and evidence-grading helpers.
irt_validation_spec( model_id, replications = 250L, parameters = NULL, metrics = c("bias", "rmse", "coverage", "interval_width", "convergence"), grouped_validation = c("device", "session", "site"), preprocessing_variants = NULL, misspecification_scenarios = NULL, thresholds = list(max_abs_bias = 0.1, max_rmse = 0.3, min_coverage = 0.9, max_failure_rate = 0.05, min_external_folds = 2L), seed = 20260808L, notes = NULL )irt_validation_spec( model_id, replications = 250L, parameters = NULL, metrics = c("bias", "rmse", "coverage", "interval_width", "convergence"), grouped_validation = c("device", "session", "site"), preprocessing_variants = NULL, misspecification_scenarios = NULL, thresholds = list(max_abs_bias = 0.1, max_rmse = 0.3, min_coverage = 0.9, max_failure_rate = 0.05, min_external_folds = 2L), seed = 20260808L, notes = NULL )
model_id |
Stable model identifier. |
replications |
Number of simulation replications planned. |
parameters |
Parameter families expected to be recovered. |
metrics |
Recovery metrics to require. |
grouped_validation |
Grouping variables for transport validation. |
preprocessing_variants |
Optional named preprocessing variants. |
misspecification_scenarios |
Optional named scenarios. |
thresholds |
Named evidence thresholds. |
seed |
Reproducibility seed. |
notes |
Free-text scientific notes. |
An object of class "eye_irt_validation_spec", stored as a named list, with components "model_id", "replications", "parameters", "metrics", "grouped_validation", "preprocessing_variants", "misspecification_scenarios", "thresholds", "seed", "notes", "contract_version", "created_with". It contains specify a validation programme for a process-IRT model and associated metadata or diagnostics needed to interpret the result.
Multi-objective item-bank optimization. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
item_objective_spec(information, process_burden, fairness, exposure, content_constraints = NULL, weights = c(information = 1, process_burden = 1, fairness = 1, exposure = 1), directions = c(information = "max", process_burden = "min", fairness = "min", exposure = "min")) item_pareto_front(x, objectives) optimize_item_bank(x, n_items, objectives, constraints = NULL, method = c("integer", "evolutionary"), iterations = 500, seed = 20260807) audit_bank_decision_stability(x, draws = 1000, noise_sd = 0.1, seed = 20260807) plot_item_pareto(x, ...) plot_objective_tradeoffs(x, ...) plot_bank_information_coverage(x, ...) plot_decision_stability(x, ...) plot_selected_bank_profile(x, ...)item_objective_spec(information, process_burden, fairness, exposure, content_constraints = NULL, weights = c(information = 1, process_burden = 1, fairness = 1, exposure = 1), directions = c(information = "max", process_burden = "min", fairness = "min", exposure = "min")) item_pareto_front(x, objectives) optimize_item_bank(x, n_items, objectives, constraints = NULL, method = c("integer", "evolutionary"), iterations = 500, seed = 20260807) audit_bank_decision_stability(x, draws = 1000, noise_sd = 0.1, seed = 20260807) plot_item_pareto(x, ...) plot_objective_tradeoffs(x, ...) plot_bank_information_coverage(x, ...) plot_decision_stability(x, ...) plot_selected_bank_profile(x, ...)
information |
Argument controlling 'information'; see the function usage and returned audit metadata. |
process_burden |
Argument controlling 'process_burden'; see the function usage and returned audit metadata. |
fairness |
Argument controlling 'fairness'; see the function usage and returned audit metadata. |
exposure |
Argument controlling 'exposure'; see the function usage and returned audit metadata. |
content_constraints |
Argument controlling 'content_constraints'; see the function usage and returned audit metadata. |
weights |
Argument controlling 'weights'; see the function usage and returned audit metadata. |
directions |
Argument controlling 'directions'; see the function usage and returned audit metadata. |
x |
Input object or data structure appropriate for the selected analysis. |
objectives |
Argument controlling 'objectives'; see the function usage and returned audit metadata. |
n_items |
Argument controlling 'n_items'; see the function usage and returned audit metadata. |
constraints |
Argument controlling 'constraints'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
iterations |
Argument controlling 'iterations'; see the function usage and returned audit metadata. |
seed |
Argument controlling 'seed'; see the function usage and returned audit metadata. |
draws |
Argument controlling 'draws'; see the function usage and returned audit metadata. |
noise_sd |
Argument controlling 'noise_sd'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Public roadmap alias for 'stress_test_latent_distribution()'.
latent_distribution_stress_test(...)latent_distribution_stress_test(...)
... |
Additional arguments passed to the selected model, engine, or method. |
An R object containing stress-test IRT estimators across latent distributions. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Estimate a descriptive continuous-time latent trajectory
latent_trait_trajectory(time, theta, spar = NULL)latent_trait_trajectory(time, theta, spar = NULL)
time |
Time values. |
theta |
Latent-trait values. |
spar |
Value supplied to 'spar'; see Details for its model-specific role. |
An object of class "eye_latent_trait_trajectory", stored as a named list, with components "model", "time", "theta". It contains a descriptive continuous-time latent trajectory and associated metadata or diagnostics needed to interpret the result.
Leave device out validation
leave_device_out_validation(data, device, fitter, predictor, scorer)leave_device_out_validation(data, device, fitter, predictor, scorer)
data |
Input data frame or compatible tabular object. |
device |
Device identifier or device facet. |
fitter |
Model-fitting function. |
predictor |
Prediction function. |
scorer |
Function that scores predictions. |
An object of class "eye_leave_device_out_validation", "data.frame", stored as a data frame, containing leave device out validation and associated metadata needed to interpret the result.
Leave item out validation
leave_item_out_validation(data, item, fitter, predictor, scorer)leave_item_out_validation(data, item, fitter, predictor, scorer)
data |
Input data frame or compatible tabular object. |
item |
Item identifier, name, or item column. |
fitter |
Model-fitting function. |
predictor |
Prediction function. |
scorer |
Function that scores predictions. |
An object of class "eye_leave_item_out_validation", "data.frame", stored as a data frame, containing leave item out validation and associated metadata needed to interpret the result.
Leave session out validation
leave_session_out_validation(data, session, fitter, predictor, scorer)leave_session_out_validation(data, session, fitter, predictor, scorer)
data |
Input data frame or compatible tabular object. |
session |
Session identifier or session facet. |
fitter |
Model-fitting function. |
predictor |
Prediction function. |
scorer |
Function that scores predictions. |
An object of class "eye_leave_session_out_validation", "data.frame", stored as a data frame, containing leave session out validation and associated metadata needed to interpret the result.
Leave site out validation
leave_site_out_validation(data, site, fitter, predictor, scorer)leave_site_out_validation(data, site, fitter, predictor, scorer)
data |
Input data frame or compatible tabular object. |
site |
Site identifier or site facet. |
fitter |
Model-fitting function. |
predictor |
Prediction function. |
scorer |
Function that scores predictions. |
An object of class "eye_leave_site_out_validation", "data.frame", stored as a data frame, containing leave site out validation and associated metadata needed to interpret the result.
List registered multimodal IRT models
list_irt_models()list_irt_models()
A tabular R object containing list registered multimodal IRT models; rows represent analysis units and columns contain the returned quantities.
Lock a decision manifest by content hash
lock_decision_manifest(x, label = "analysis_decisions")lock_decision_manifest(x, label = "analysis_decisions")
x |
Manifest. |
label |
Optional lock label. |
An object of class "eye_decision_manifest_lock", stored as a named list, with components "manifest", "manifest_hash", "label", "locked_at", "status". It contains lock a decision manifest by content hash and associated metadata or diagnostics needed to interpret the result.
Map supplied latent-class memberships to process summaries
map_latent_classes_to_process_profiles( class_membership, process_data, person = "person_id", class_col = "class", process_features )map_latent_classes_to_process_profiles( class_membership, process_data, person = "person_id", class_col = "class", process_features )
class_membership |
Data containing person/class assignments or probabilities. |
process_data |
Person-level or trial-level process data. |
person |
Person identifier present in both objects. |
class_col |
Class-assignment column. |
process_features |
Numeric process features. |
An object of class "eye_latent_process_alignment", stored as a named list, with components "data", "summary", "process_features", "class_col", "caveat". It contains supplied latent-class memberships to process summaries and associated metadata or diagnostics needed to interpret the result.
Build a non-collapsed measurement-error budget
measurement_error_budget( accuracy = NA_real_, precision = NA_real_, data_loss = NA_real_, effective_hz = NA_real_, calibration_drift = NA_real_, units = NULL )measurement_error_budget( accuracy = NA_real_, precision = NA_real_, data_loss = NA_real_, effective_hz = NA_real_, calibration_drift = NA_real_, units = NULL )
accuracy |
Measurement inaccuracy/offset metric. |
precision |
Measurement imprecision metric. |
data_loss |
Data-loss proportion. |
effective_hz |
Effective sampling frequency. |
calibration_drift |
Optional drift metric. |
units |
Optional named units. |
A data frame containing a non-collapsed measurement-error budget. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
migrate_eye_storage_schema(storage, target_path, target_version = "2.0.0", format = NULL, overwrite = FALSE)migrate_eye_storage_schema(storage, target_path, target_version = "2.0.0", format = NULL, overwrite = FALSE)
storage |
Storage object or path. |
target_path |
Destination. |
target_version |
Target schema version. |
format |
Target format. |
overwrite |
Whether to replace target. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
model_promotion_spec(model_families = c("dynamic_irtree", "functional_pupil_irt", "theory_strategy_irt", "gaze_diffusion_irt"), require_completion = TRUE, require_sbc = TRUE, require_misspecification = TRUE, require_grouped_validation = TRUE, require_engine_equivalence = TRUE, require_empirical_reproduction = TRUE, require_preprocessing_sensitivity = TRUE, require_multi_vendor = FALSE)model_promotion_spec(model_families = c("dynamic_irtree", "functional_pupil_irt", "theory_strategy_irt", "gaze_diffusion_irt"), require_completion = TRUE, require_sbc = TRUE, require_misspecification = TRUE, require_grouped_validation = TRUE, require_engine_equivalence = TRUE, require_empirical_reproduction = TRUE, require_preprocessing_sensitivity = TRUE, require_multi_vendor = FALSE)
model_families |
Model families to audit. |
require_completion |
Require complete Monte Carlo evidence. |
require_sbc |
Require simulation-based calibration. |
require_misspecification |
Require declared misspecification studies. |
require_grouped_validation |
Require grouped out-of-sample validation. |
require_engine_equivalence |
Require external-engine comparison. |
require_empirical_reproduction |
Require an empirical reproduction. |
require_preprocessing_sensitivity |
Require preprocessing/AOI sensitivity. |
require_multi_vendor |
Require independent multi-vendor evidence. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Specify a model-validation programme
model_validation_spec( replications = 100L, confidence = 0.95, max_abs_bias = 0.1, min_coverage = 0.9, max_failure_rate = 0.05 )model_validation_spec( replications = 100L, confidence = 0.95, max_abs_bias = 0.1, min_coverage = 0.9, max_failure_rate = 0.05 )
replications |
Number of Monte Carlo replications. |
confidence |
Confidence level for interval coverage. |
max_abs_bias |
Maximum acceptable absolute bias. |
min_coverage |
Minimum acceptable interval coverage. |
max_failure_rate |
Maximum acceptable estimation failure rate. |
An 'eye_model_validation_spec'.
Summarize model validation
model_validation_summary(x)model_validation_summary(x)
x |
An 'eye_model_validation' object. |
Scenario-by-parameter validation metrics.
Extract multiblock block contributions/coordinates
multiblock_contributions(x)multiblock_contributions(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing multiblock block contributions/coordinates. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Extract multiblock person coordinates
multiblock_person_coordinates(x)multiblock_person_coordinates(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing multiblock person coordinates. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Extract multiblock variable coordinates
multiblock_variable_coordinates(x)multiblock_variable_coordinates(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing multiblock variable coordinates. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Vendor-neutral preparation and auditing utilities for repeated-measures mediation. The functions preserve trial-level rows, separate within- and between-participant variation, distinguish observed zero mediators from unobserved mediators, and retain provenance needed by downstream specialist estimators.
center_within_participant( data, value_col, participant_col = "participant_id", output_col = NULL ) decompose_within_between( data, columns, participant_col = "participant_id", grand_mean_center_between = FALSE ) summarise_within_between_variance( data, columns, participant_col = "participant_id" ) identify_mediation_levels( data, columns, participant_col = "participant_id", tolerance = 1e-12 ) audit_mediation_missingness( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", quality_col = NULL, minimum_quality = NULL, mediator_observed_col = NULL, response_observed_col = NULL ) check_mediation_trial_counts( data, participant_col = "participant_id", trial_col = "trial_id", minimum_trials = 2L ) validate_multilevel_mediation_data( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", require_within_x = TRUE ) prepare_multilevel_mediation_data( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", quality_col = NULL, minimum_quality = NULL, mediator_observed_col = NULL, response_observed_col = NULL, quality_action = c("flag", "mask_mediator"), grand_mean_center_between = FALSE, require_within_x = TRUE, source_id = NULL, preprocessing_spec = NULL, event_detector = NULL, aoi_specification = NULL, warn = TRUE ) add_multilevel_mediation_component( prepared, value_col, semantic, within_col, between_col, grand_mean_center_between = NULL )center_within_participant( data, value_col, participant_col = "participant_id", output_col = NULL ) decompose_within_between( data, columns, participant_col = "participant_id", grand_mean_center_between = FALSE ) summarise_within_between_variance( data, columns, participant_col = "participant_id" ) identify_mediation_levels( data, columns, participant_col = "participant_id", tolerance = 1e-12 ) audit_mediation_missingness( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", quality_col = NULL, minimum_quality = NULL, mediator_observed_col = NULL, response_observed_col = NULL ) check_mediation_trial_counts( data, participant_col = "participant_id", trial_col = "trial_id", minimum_trials = 2L ) validate_multilevel_mediation_data( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", require_within_x = TRUE ) prepare_multilevel_mediation_data( data, x_col, mediator_col, outcome_col, participant_col = "participant_id", trial_col = "trial_id", quality_col = NULL, minimum_quality = NULL, mediator_observed_col = NULL, response_observed_col = NULL, quality_action = c("flag", "mask_mediator"), grand_mean_center_between = FALSE, require_within_x = TRUE, source_id = NULL, preprocessing_spec = NULL, event_detector = NULL, aoi_specification = NULL, warn = TRUE ) add_multilevel_mediation_component( prepared, value_col, semantic, within_col, between_col, grand_mean_center_between = NULL )
data |
Repeated-measures trial-level data frame. Rows are preserved. |
value_col |
Numeric variable to decompose or add as a mediation component. |
columns |
One or more numeric variables to inspect or decompose. |
participant_col |
Participant identifier column. |
trial_col |
Trial identifier column. |
output_col |
Optional name for a within-participant centered variable. |
grand_mean_center_between |
Whether participant means are centered around the observed grand mean. |
tolerance |
Non-negative tolerance for classifying within/between variation. |
x_col |
Exposure or manipulation column. |
mediator_col |
Trial-level mediator column. |
outcome_col |
Trial-level outcome column. |
quality_col |
Optional numeric trial-quality column. |
minimum_quality |
Required finite threshold when |
mediator_observed_col |
Optional explicit marker that the mediator was observed. |
response_observed_col |
Optional explicit marker that the outcome was observed. |
minimum_trials |
Minimum trial count used for the descriptive participant audit. |
require_within_x |
Whether preparation requires within-participant exposure variation. |
quality_action |
Either |
source_id |
Optional source-data identifier retained in provenance. |
preprocessing_spec, event_detector, aoi_specification
|
Optional preprocessing, event-detector, and AOI provenance. |
warn |
Whether preparation emits informative warnings stored in the returned object. |
prepared |
An |
semantic |
Semantic name used to register an additional mediation component. |
within_col, between_col
|
Explicit output columns for an added within/between component. |
The preparation layer does not fit a mediation estimator and does not convert
missing mediators to zero. Quality rules are explicit. When
quality_action = "mask_mediator", masking occurs before decomposition and
is recorded in provenance. Within- and between-participant components are kept
separate so a downstream multilevel estimator can model the intended level
rather than conflating trial and person variation.
Centering and decomposition functions return data frames. Audit functions return
data frames or validation objects. prepare_multilevel_mediation_data()
returns an eye_multilevel_mediation_data object containing the preserved
trial table, audits, column mappings, and provenance.
d <- data.frame( participant_id = rep(c("P1", "P2"), each = 3), trial_id = paste0("T", 1:6), condition = c(0, 1, 0, 1, 0, 1), dwell = c(1.0, 1.4, 1.2, 0.8, 1.1, 1.3), outcome = c(3, 4, 3, 2, 3, 4) ) x <- prepare_multilevel_mediation_data( d, x_col = "condition", mediator_col = "dwell", outcome_col = "outcome" ) head(x$data[c("X_within", "X_between", "M_within", "M_between")])d <- data.frame( participant_id = rep(c("P1", "P2"), each = 3), trial_id = paste0("T", 1:6), condition = c(0, 1, 0, 1, 0, 1), dwell = c(1.0, 1.4, 1.2, 0.8, 1.1, 1.3), outcome = c(3, 4, 3, 2, 3, 4) ) x <- prepare_multilevel_mediation_data( d, x_col = "condition", mediator_col = "dwell", outcome_col = "outcome" ) head(x$data[c("X_within", "X_between", "M_within", "M_between")])
Report multimodal backend availability
multimodal_backend_status()multimodal_backend_status()
A data.frame describing optional backend availability.
This is a thin multimodal adapter over the established 'irt_model_spec()' / 'eye_irt_model_spec' architecture. It composes existing 'eye_irt_channel' objects and does not define a parallel channel or model-specification ecosystem.
multimodal_irt_spec( response = NULL, rt = NULL, gaze = NULL, pupil = NULL, model = c("M0", "M1", "M2", "M3"), backend = c("cmdstanr", "existing"), identification = list(), priors = list() )multimodal_irt_spec( response = NULL, rt = NULL, gaze = NULL, pupil = NULL, model = c("M0", "M1", "M2", "M3"), backend = c("cmdstanr", "existing"), identification = list(), priors = list() )
response, rt, gaze, pupil
|
Existing 'eye_irt_channel' objects or NULL. |
model |
Development model identifier. |
backend |
Requested backend. |
identification |
Named identification settings retained as explicit multimodal metadata. |
priors |
Named prior settings retained as explicit multimodal metadata. |
An 'eye_multimodal_irt_spec' convenience subclass of the established 'eye_irt_model_spec'.
Fits response-only (M0), response+RT (M1), and response+RT+gaze (M2) with compatible hierarchical Stan implementations. The sequence is designed for response-target comparison rather than for asserting that information from distinct channels is algebraically additive.
multimodal_m2_ablation(x, ...)multimodal_m2_ablation(x, ...)
x |
Data accepted by [fit_multimodal_m2()]. |
... |
Sampling arguments forwarded to the internal reference fitters. |
An 'eye_multimodal_m2_ablation'.
Generates deterministic within-item permutations that preserve each item's marginal channel distribution while breaking person-level alignment for gaze, RT, or response. These are falsification controls, not causal interventions and not misconduct detectors.
multimodal_m2_negative_controls(x, seed = 20260814L)multimodal_m2_negative_controls(x, seed = 20260814L)
x |
M2-compatible data. |
seed |
Seed for deterministic permutations. |
An 'eye_multimodal_m2_negative_controls'.
Reproduces the channel-specific discrepancy logic used in the published three-way framework: response W, response-time L, and fixation-count M residual statistics, aggregated by item.
multimodal_m2_ppc(x)multimodal_m2_ppc(x)
x |
An 'eye_multimodal_m2_fit' with 'model == "M2"'. |
An 'eye_multimodal_m2_ppc'.
Computes two complementary quantities on a common response target: response-target PSIS-LOO ELPD and posterior variance of person ability. This avoids simply summing channel Fisher information under a joint correlated model.
multimodal_m2_process_information(x)multimodal_m2_process_information(x)
x |
An 'eye_multimodal_m2_ablation'. |
An 'eye_multimodal_m2_information'.
Repeatedly simulates from the M2 generating model, fits the M2 CmdStan reference estimator, and computes bias, RMSE, posterior standard deviation, and 95 dispersion, correlation, and hyperparameter families.
multimodal_m2_recovery( n_rep = 10L, n_person = 100L, n_item = 10L, dropout = c(response = 0, rt = 0, gaze = 0), base_seed = 20260814L, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, prior_profile = c("regularized", "paper_centered"), adapt_delta = 0.95, max_treedepth = 12L, refresh = 0L )multimodal_m2_recovery( n_rep = 10L, n_person = 100L, n_item = 10L, dropout = c(response = 0, rt = 0, gaze = 0), base_seed = 20260814L, chains = 4L, parallel_chains = chains, iter_warmup = 1000L, iter_sampling = 1000L, prior_profile = c("regularized", "paper_centered"), adapt_delta = 0.95, max_treedepth = 12L, refresh = 0L )
n_rep |
Number of simulation/fit replications. |
n_person, n_item
|
Simulation size. |
dropout |
Channel dropout probabilities. |
base_seed |
Base seed. |
chains, parallel_chains, iter_warmup, iter_sampling
|
CmdStan controls. |
prior_profile |
Prior profile. |
adapt_delta, max_treedepth, refresh
|
CmdStan controls. |
This is intentionally computationally expensive and is not executed during ordinary package tests.
An 'eye_multimodal_m2_recovery'.
Creates a thin M2 specialization of the established [multimodal_irt_spec()] / 'irt_model_spec()' architecture. The measurement likelihood follows Man, Harring, and Zhan (2022): Rasch response, lognormal response time, and negative-binomial gaze-fixation counts. Priors are implemented in Stan using either a regularized profile or a paper-centered profile; the latter is not claimed to reproduce every published hyperprior exactly.
multimodal_m2_spec( backend = "cmdstanr", prior_profile = c("regularized", "paper_centered"), missingness = "ignorable" )multimodal_m2_spec( backend = "cmdstanr", prior_profile = c("regularized", "paper_centered"), missingness = "ignorable" )
backend |
Currently '"cmdstanr"' only. |
prior_profile |
'"regularized"' or '"paper_centered"'. |
missingness |
Currently '"ignorable"' only. Channel-specific missing observations contribute no level-1 likelihood term. |
An 'eye_multimodal_m2_spec', inheriting the established 'eye_multimodal_irt_spec' and 'eye_irt_model_spec' classes.
Fits all eight response-anchored combinations of RT, gaze and pupil. Models use one common subset-likelihood implementation to make the target and prior family explicit. This is inferential ablation, not a causal intervention on sensors.
multimodal_m3_ablation( x, models = NULL, nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), ... )multimodal_m3_ablation( x, models = NULL, nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), ... )
x |
M3-compatible data. |
models |
Optional subset of the eight model identifiers. |
nuisance |
Named logical vector selecting pupil nuisance terms. The same selection is applied to every ablation model containing pupil. |
... |
Sampling controls forwarded to the subset fitter. |
An 'eye_multimodal_m3_ablation'.
This bridge deliberately requires an analyst-supplied, already-derived trial-level functional score. It does not silently reduce a raw time series. Existing 'functional_pupil_irt_spec()', deconvolution, and confound tools remain the authoritative trajectory-level machinery.
multimodal_m3_functional_bridge( data, score, pupil = "pupil", provenance = NULL )multimodal_m3_functional_bridge( data, score, pupil = "pupil", provenance = NULL )
data |
Trial-level M3 data. |
score |
Trial-level functional/trajectory score or its column name. |
pupil |
Name of the output pupil column. |
provenance |
Free-text derivation/provenance note. |
An 'eye_multimodal_m3_functional_bridge' data object.
Includes within-item RT/gaze/pupil permutations, within-person pupil permutation across items, pupil phase randomization, luminance-only pseudo-pupil, and an irrelevant synthetic channel. Controls preserve selected marginals while deliberately breaking alignment; they are not causal interventions or misconduct detectors.
multimodal_m3_negative_controls(x, seed = 20260815L)multimodal_m3_negative_controls(x, seed = 20260815L)
x |
M3-compatible data. |
seed |
Deterministic seed. |
An 'eye_multimodal_m3_negative_controls'.
Extends the M2 W/L/M discrepancy family with a standardized pupil residual discrepancy P. Tail flags are diagnostics, not construct validation.
multimodal_m3_ppc(x)multimodal_m3_ppc(x)
x |
An M3 fit. |
An 'eye_multimodal_m3_ppc'.
Uses response-target PSIS-LOO and posterior ability variance across the eight-channel ablation lattice. It additionally reports paired pupil gains, a non-additivity/redundancy contrast, a channel-conflict screen, and optional information per usable pupil observation or sensor cost. Positive values do not establish construct validity or causal sensor value.
multimodal_m3_process_information(x, pupil_cost = 1, decisive_z = 2)multimodal_m3_process_information(x, pupil_cost = 1, decisive_z = 2)
x |
An M3 ablation object. |
pupil_cost |
Optional positive relative pupil-sensor cost. |
decisive_z |
Absolute delta/SE ratio used only as a descriptive evidence flag. |
An 'eye_multimodal_m3_information' object.
Repeatedly simulates and fits M3 across pupil-signal and pupil-missingness scenarios, summarizing bias, RMSE, posterior SD, interval coverage and MCMC diagnostics. Small 'reps' values are smoke tests; scientific promotion requires a predeclared larger grid.
multimodal_m3_recovery( reps = 3L, pupil_signal = c("informative", "weak", "null", "redundant", "confounded"), pupil_missingness = c("mcar", "quality", "device"), n_person = 80L, n_item = 10L, seed = 20260815L, fit_args = list(chains = 2L, parallel_chains = 2L, iter_warmup = 500L, iter_sampling = 300L, refresh = 0L, init = 0) )multimodal_m3_recovery( reps = 3L, pupil_signal = c("informative", "weak", "null", "redundant", "confounded"), pupil_missingness = c("mcar", "quality", "device"), n_person = 80L, n_item = 10L, seed = 20260815L, fit_args = list(chains = 2L, parallel_chains = 2L, iter_warmup = 500L, iter_sampling = 300L, refresh = 0L, init = 0) )
reps |
Replications per design cell. |
pupil_signal, pupil_missingness
|
Scenario vectors. |
n_person, n_item
|
Design size. |
seed |
Base seed. |
fit_args |
Named sampling arguments for 'fit_multimodal_m3()'. |
An 'eye_multimodal_m3_recovery' object.
Extends the established eyeprocess multimodal IRT specification with a continuous pupil-responsivity channel. The summary representation is fitted by the bundled four-channel Stan model. 'functional_score' records that the pupil input was derived from a trajectory/functional workflow; it does not silently turn the scalar reference likelihood into a functional likelihood.
multimodal_m3_spec( backend = "cmdstanr", prior_profile = c("regularized", "paper_centered"), missingness = "ignorable", pupil_representation = c("summary", "functional_score"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names) )multimodal_m3_spec( backend = "cmdstanr", prior_profile = c("regularized", "paper_centered"), missingness = "ignorable", pupil_representation = c("summary", "functional_score"), nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names) )
backend |
Currently '"cmdstanr"' only. |
prior_profile |
Prior profile. |
missingness |
Currently '"ignorable"' only. |
pupil_representation |
'"summary"' or '"functional_score"'. |
nuisance |
Named logical vector selecting baseline, luminance, gaze X/Y, quality, blink, interpolation and time-on-task nuisance terms. Availability is also audited from data. |
An 'eye_multimodal_m3_spec' inheriting the established multimodal and IRT spec classes.
The default is plan-only. The essential lattice compares M3, formal K=1, reference K=2, K=2 without trait conditioning, and K=2 with iid states. Optional RT-anchored channel ablations are available but are not part of the default development burden. A no-RT K>1 model is intentionally excluded because the reference label-identification policy orders RT state deviations.
multimodal_m4_ablation( x, run = FALSE, include_channel_ablations = FALSE, m3_fit = NULL, m4_fit = NULL, fit_args = list() )multimodal_m4_ablation( x, run = FALSE, include_channel_ablations = FALSE, m3_fit = NULL, m4_fit = NULL, fit_args = list() )
x |
Data or M4 simulation. |
run |
Whether to fit models. Default 'FALSE'. |
include_channel_ablations |
Add RT-only, no-gaze, and no-pupil state models. |
m3_fit, m4_fit
|
Optional already fitted baseline/reference objects to reuse. |
fit_args |
Sampling arguments shared across new fits. |
An 'eye_multimodal_m4_ablation'.
Controls are transformations/stress designs by default and therefore do not incur backend fitting. Set 'run = TRUE' only when explicit fitted comparisons are needed.
multimodal_m4_negative_controls( x, controls = c("order_shuffle", "process_shuffle", "state_independent", "nuisance_pseudostate", "device_session_pseudostate", "overfit_state_count"), seed = 20260820L, run = FALSE, fit_args = list() )multimodal_m4_negative_controls( x, controls = c("order_shuffle", "process_shuffle", "state_independent", "nuisance_pseudostate", "device_session_pseudostate", "overfit_state_count"), seed = 20260820L, run = FALSE, fit_args = list() )
x |
Data or M4 simulation. |
controls |
Control names. |
seed |
Deterministic seed. |
run |
Whether to fit the generated controls. |
fit_args |
Arguments for optional M4 fits. |
An 'eye_multimodal_m4_negative_controls'.
Compares observed process summaries with posterior-replicated summaries and reports replicated latent-state dynamics separately. Because states are latent, inferred state labels are never treated as observed ground truth.
multimodal_m4_ppc(x)multimodal_m4_ppc(x)
x |
An M4 fit. |
An 'eye_multimodal_m4_ppc'.
Uses commensurate response-target PSIS-LOO and ability-posterior uncertainty to compare M3 with focused M4 variants. It explicitly allows no benefit, redundancy, or destabilization and does not sum channel Fisher information.
multimodal_m4_process_information(x, decisive_z = 2)multimodal_m4_process_information(x, decisive_z = 2)
x |
Executed M4 ablation object. |
decisive_z |
Descriptive absolute delta/SE threshold. |
An 'eye_multimodal_m4_information'.
Recovery aligns latent-state labels before scoring multichannel state effects, and emphasizes posterior-probability calibration, occupancy, and transition recovery rather than treating MAP classification accuracy as the primary criterion. By default the function returns a small five-scenario design and does not launch expensive fitting.
multimodal_m4_recovery( simulation = NULL, fit = NULL, scenarios = c("clear", "weak", "null", "trait_conditioned", "nuisance_confounded"), run = FALSE, simulation_args = list(n_person = 60L, n_item = 10L), fit_args = list() )multimodal_m4_recovery( simulation = NULL, fit = NULL, scenarios = c("clear", "weak", "null", "trait_conditioned", "nuisance_confounded"), run = FALSE, simulation_args = list(n_person = 60L, n_item = 10L), fit_args = list() )
simulation |
Optional single M4 simulation. |
fit |
Optional already-fitted M4 model corresponding to 'simulation'. |
scenarios |
Deterministic development battery used when 'run = TRUE'. |
run |
Whether to execute the small recovery battery. Default 'FALSE'. |
simulation_args |
Named arguments forwarded to simulation. |
fit_args |
Named arguments forwarded to 'fit_multimodal_m4()'. |
An 'eye_multimodal_m4_recovery'.
The default returns a transparent sensitivity design without fitting. It covers K, priors, trait conditioning, transition structure, state channels, nuisance adjustment, and sequence-length threshold. No automatic 'best_K' is declared.
multimodal_m4_sensitivity(x, n_states = 1:4, run = FALSE, fit_args = list())multimodal_m4_sensitivity(x, n_states = 1:4, run = FALSE, fit_args = list())
x |
Data or simulation. |
n_states |
Candidate state counts, default 1:4. |
run |
Whether to execute K-sensitivity fits. Other sensitivity dimensions remain explicit design rows rather than an automatic combinatorial grid. |
fit_args |
Optional fitting arguments. |
An 'eye_multimodal_m4_sensitivity'.
Defines the M4 sequential extension of the validated M3 response + RT + gaze + pupil model. The reference implementation keeps the scored-response Rasch equation state-independent and lets latent states shift selected RT, gaze, and pupil process channels. State labels are statistical identification labels only and do not imply psychological constructs.
multimodal_m4_spec( n_states = 2L, state_channels = c("rt", "gaze", "pupil"), transition_structure = c("markov", "iid"), trait_conditioning = c("theta", "tau"), initial_trait_conditioning = TRUE, min_sequence_length = 2L, identification = c("ordered_rt_effect"), prior_profile = c("regularized", "paper_centered"), missingness = "ignorable", nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), backend = "cmdstanr" )multimodal_m4_spec( n_states = 2L, state_channels = c("rt", "gaze", "pupil"), transition_structure = c("markov", "iid"), trait_conditioning = c("theta", "tau"), initial_trait_conditioning = TRUE, min_sequence_length = 2L, identification = c("ordered_rt_effect"), prior_profile = c("regularized", "paper_centered"), missingness = "ignorable", nuisance = stats::setNames(rep(TRUE, 8L), .ep10_m3_nuisance_names), backend = "cmdstanr" )
n_states |
Number of latent states, from 1 through 4. '1' is the formal null state model and should be treated as scientifically meaningful. |
state_channels |
Subset of '"rt"', '"gaze"', and '"pupil"' receiving state-dependent deviations. |
transition_structure |
'"markov"' for first-order transitions or '"iid"' for independent state membership over ordered trials. |
trait_conditioning |
Subset of 'theta', 'tau', 'omega', and 'rho' used to condition transition logits. The conservative default is 'theta + tau'. |
initial_trait_conditioning |
Whether the same selected traits condition initial-state probabilities. |
min_sequence_length |
Minimum sequence length required by the structural audit for fitting. No rows are silently removed when sequences are shorter. |
identification |
State-label identification policy. The reference policy orders centered RT state deviations; this is a label convention only. |
prior_profile |
Prior profile inherited from M3. |
missingness |
Currently '"ignorable"' only. |
nuisance |
Named pupil-nuisance selection vector inherited from M3. |
backend |
Currently '"cmdstanr"' only. |
An 'eye_multimodal_m4_spec' inheriting the canonical eyeprocess multimodal/IRT specification classes.
Returns posterior state probabilities, entropy, occupancy, MAP-run summaries, transition probabilities, and posterior-weighted process-channel profiles. MAP states are explicitly secondary summaries of posterior probabilities.
multimodal_m4_state_diagnostics(x)multimodal_m4_state_diagnostics(x)
x |
M4 fit or M4 simulation. Simulation diagnostics use known synthetic truth and are labelled accordingly. |
An 'eye_multimodal_m4_states' object.
Posterior predictive checks for multimodal development fits
multimodal_ppc(object, variables = NULL)multimodal_ppc(object, variables = NULL)
object |
An 'eye_multimodal_irt_fit'. |
variables |
Optional generated-quantity variable names. |
An 'eye_multimodal_ppc'.
Concordance of multiple negative-control families
negative_control_concordance(x, effect = "effect", tolerance = 0.05)negative_control_concordance(x, effect = "effect", tolerance = 0.05)
x |
Negative-control result. |
effect |
Effect column. |
tolerance |
Absolute mean-effect tolerance. |
A named list with components "summary", "all_within_tolerance", "tolerance", containing concordance of multiple negative-control families and associated metadata or diagnostics.
The user supplies a complete evaluation callback. eyeprocess permutes the named process columns, optionally within grouping strata, and compares the observed score with the permutation distribution.
negative_control_process_test( data, process_columns, evaluator, within = NULL, permutations = 100L, higher_is_better = TRUE, seed = 20260808L )negative_control_process_test( data, process_columns, evaluator, within = NULL, permutations = 100L, higher_is_better = TRUE, seed = 20260808L )
data |
Input data. |
process_columns |
Columns to permute. |
evaluator |
Function returning one scalar out-of-sample score. |
within |
Optional grouping columns within which permutation occurs. |
permutations |
Number of negative-control permutations. |
higher_is_better |
Score direction. |
seed |
Random-number seed. |
An object of class "eye_process_negative_control", stored as a named list, with components "observed", "null", "p_value", "permutations", "process_columns", "within", "higher_is_better", "seed". It contains negative-control test for an allegedly informative process channel and associated metadata or diagnostics needed to interpret the result.
Hash an R object reproducibly within an R serialization version
object_hash(x)object_hash(x)
x |
R object. |
An R object containing hash an R object reproducibly within an R serialization version. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
object_schema(object = c("eye_dataset", "eyeprocess_model", "validation_plan", "validation_collection", "vendor_corpus", "eye_storage"))object_schema(object = c("eye_dataset", "eyeprocess_model", "validation_plan", "validation_collection", "vendor_corpus", "eye_storage"))
object |
Object or schema name. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Open an eyeprocess storage handle
open_eye_storage(path, format = NULL)open_eye_storage(path, format = NULL)
path |
Storage path. |
format |
Optional explicit format. |
An 'eye_storage' handle without collecting all tables.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
open_partitioned_eye_storage(path)open_partitioned_eye_storage(path)
path |
Storage directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Quantify option-process information from a nominal gaze model
option_process_information(object)option_process_information(object)
object |
A fitted eyeprocess model or audit object. |
An object of class "eye_option_process_information", "data.frame", stored as a data frame, containing quantify option-process information from a nominal gaze model and associated metadata needed to interpret the result.
Audit whether candidate predictors can be constructed without an outcome column
outcome_blind_feature_audit(data, outcome, feature_fun)outcome_blind_feature_audit(data, outcome, feature_fun)
data |
Data frame. |
outcome |
Outcome column name. |
feature_fun |
Function receiving outcome-hidden data and returning features. |
A named list with components "status", "outcome", "feature_result", "error", "warnings", "input_columns", "interpretation", containing whether candidate predictors can be constructed without an outcome column and associated metadata or diagnostics.
Create an outcome-blind data snapshot
outcome_blind_snapshot(data, outcome, id = NULL)outcome_blind_snapshot(data, outcome, id = NULL)
data |
Data frame. |
outcome |
Outcome column(s) to remove from the analysis snapshot. |
id |
Optional identifier columns retained in the snapshot. |
An object of class "eye_outcome_blind_snapshot", stored as a named list, with components "data", "removed_outcomes", "id", "source_columns", "blinded_columns", "blinded_hash", "created_at", "caveat". It contains an outcome-blind data snapshot and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
package_reproducibility_manifest(paths, include_session = TRUE)package_reproducibility_manifest(paths, include_session = TRUE)
paths |
Files/directories. |
include_session |
Whether to capture session information. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Create a compact paper reproducibility manifest
paper_reproducibility_manifest( evidence, manuscript = NULL, figures = NULL, tables = NULL )paper_reproducibility_manifest( evidence, manuscript = NULL, figures = NULL, tables = NULL )
evidence |
Evidence bundle. |
manuscript |
Optional manuscript path. |
figures |
Optional figure paths. |
tables |
Optional table paths. |
A named list with components "evidence_hash", "manuscript", "files", "reproducibility", "generated_utc", containing a compact paper reproducibility manifest and associated metadata or diagnostics.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
partition_eye_storage(by = c("participant_id", "session_id", "recording_id"), format = c("parquet", "csv", "rds"), compression = "zstd", max_rows = 1000000L)partition_eye_storage(by = c("participant_id", "session_id", "recording_id"), format = c("parquet", "csv", "rds"), compression = "zstd", max_rows = 1000000L)
by |
Partition columns. |
format |
Storage format. |
compression |
Compression codec for Parquet. |
max_rows |
Maximum rows per physical file. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Return failed pipeline steps
pipeline_failures(x)pipeline_failures(x)
x |
Pipeline run. |
A data frame containing return failed pipeline steps. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Extract a pipeline result by step name
pipeline_result(x, step)pipeline_result(x, step)
x |
Pipeline run. |
step |
Step name. |
An R object containing a pipeline result by step name. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Pipeline step status table
pipeline_step_status(x)pipeline_step_status(x)
x |
Pipeline run. |
A data frame containing pipeline step status table. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Audit a placebo/pre-event window
placebo_window_audit(data, time, value, window, expected = 0, by = NULL)placebo_window_audit(data, time, value, window, expected = 0, by = NULL)
data |
Data frame. |
time |
Time column. |
value |
Value column. |
window |
Two-element placebo window. |
expected |
Optional expected mean, typically zero. |
by |
Optional grouping columns. |
An R object containing a placebo/pre-event window. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot an AOI transition matrix
plot_aoi_transition_matrix( data, from = "from", to = "to", normalize = c("from", "all", "none"), ... )plot_aoi_transition_matrix( data, from = "from", to = "to", normalize = c("from", "all", "none"), ... )
data |
Transition-pair data. |
from, to
|
Column names. |
normalize |
Normalize within from-AOI, globally, or not at all. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot an AOI transition matrix. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot top AOI transitions by probability/count
plot_aoi_transition_rank( data, from = "from", to = "to", normalize = c("from", "all", "none"), top_n = 20L, ... )plot_aoi_transition_rank( data, from = "from", to = "to", normalize = c("from", "all", "none"), top_n = 20L, ... )
data |
Data frame containing the required process variables. |
from |
Name of the column identifying the transition origin. |
to |
Name of the column identifying the transition destination. |
normalize |
Normalization rule applied to transition counts or weights. |
top_n |
Maximum number of highest-ranked entries to display. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot top AOI transitions by probability/count. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot option-level distractor information
plot_distractor_information(object, ...)plot_distractor_information(object, ...)
object |
An 'eye_nominal_gaze_irt' object or a data frame returned by 'distractor_process_map()'. |
... |
Graphical arguments passed to base graphics. |
The plotted distractor map, invisibly.
Generate the complete Gazepoint workflow plot suite
plot_gazepoint_workflow(x, directory, channels = NULL, expected_hz = 60)plot_gazepoint_workflow(x, directory, channels = NULL, expected_hz = 60)
x |
A processed 'eye_dataset'. |
directory |
Destination directory. |
channels |
Biometric channels to plot. |
expected_hz |
Expected gaze sampling rate shown in diagnostics. |
A plot manifest data frame.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
plot_interval_coverage(x, target = 0.95, engine = c("auto", "ggplot2", "base"), ...)plot_interval_coverage(x, target = 0.95, engine = c("auto", "ggplot2", "base"), ...)
x |
Validation collection or recovery summary. |
target |
Nominal coverage target. |
engine |
Plot engine. |
... |
Additional plotting arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
For the bundled spline-reference engine, standard errors are derived on the logit scale from the fitted GLM and transformed to response probabilities. Exact GPIRT engines should supply their own posterior uncertainty summaries.
plot_irf_uncertainty( object, item = 1L, theta_grid = seq(-4, 4, length.out = 101), level = 0.95, ... )plot_irf_uncertainty( object, item = 1L, theta_grid = seq(-4, 4, length.out = 101), level = 0.95, ... )
object |
An 'eye_gpirt' object. |
item |
Item name or index. |
theta_grid |
Trait grid. |
level |
Pointwise confidence level for the spline-reference diagnostic. |
... |
Graphical arguments. |
A data frame containing plot uncertainty for flexible item response functions. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
plot_parameter_recovery(x, parameter = NULL, engine = c("auto", "ggplot2", "base"), ...)plot_parameter_recovery(x, parameter = NULL, engine = c("auto", "ggplot2", "base"), ...)
x |
Validation collection or estimates data frame. |
parameter |
Optional parameter subset. |
engine |
Plot engine. |
... |
Additional plotting arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot person/item latent-space coordinates
plot_person_item_space(object, dimensions = c(1L, 2L), labels = FALSE, ...)plot_person_item_space(object, dimensions = c(1L, 2L), labels = FALSE, ...)
object |
Fitted 'eye_latent_space_irt' object. |
dimensions |
Two coordinate dimensions to display. |
labels |
Whether to add entity labels. |
... |
Graphical arguments. |
A named list with components "person", "item", "dimensions", containing plot person/item latent-space coordinates and associated metadata or diagnostics.
Plot detected process change points
plot_process_changepoint(object, ...)plot_process_changepoint(object, ...)
object |
An 'eye_irt_changepoints', 'eye_changepoint_rt_irt', or 'eye_changepoint_multimodal_irt' object. |
... |
Graphical arguments. |
An R object containing plot detected process change points. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot channel-ablation delta from a full/reference model
plot_process_channel_ablation_delta( x = NULL, table = NULL, channel_col = "channel", metric_col = "metric", value_col = "value", full_label = "full", metric = NULL, ... )plot_process_channel_ablation_delta( x = NULL, table = NULL, channel_col = "channel", metric_col = "metric", value_col = "value", full_label = "full", metric = NULL, ... )
x |
'eye_process_channel_ablation' object or compatible table. |
table |
Optional explicit table. |
channel_col, metric_col, value_col
|
Column names. |
full_label |
Full/reference channel label. |
metric |
Metric to evaluate or display. |
... |
Additional arguments passed to the underlying method or helper. |
A data frame containing plot channel-ablation delta from a full/reference model. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Plot process-feature stability across resamples/splits
plot_process_feature_stability( data, feature = "feature", stability = "selection_rate", top_n = 20L, ... )plot_process_feature_stability( data, feature = "feature", stability = "selection_rate", top_n = 20L, ... )
data |
Table with feature and stability/rank information. |
feature |
Feature column. |
stability |
Stability/selection-rate column. |
top_n |
Maximum features shown. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot process-feature stability across resamples/splits. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Explicit wrapper for process-window sensitivity plotting
plot_process_window_sensitivity(x, ...)plot_process_window_sensitivity(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing explicit wrapper for process-window sensitivity plotting. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot pupil activity sensitivity to window length
plot_pupil_activity_sensitivity(x, feature = "pupil_frequency_contrast", ...)plot_pupil_activity_sensitivity(x, feature = "pupil_frequency_contrast", ...)
x |
'eye_pupil_frequency_stability' object. |
feature |
Feature name. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot pupil activity sensitivity to window length. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot pupil activity features across windows/groups
plot_pupil_activity_windows(x, feature = "pupil_frequency_contrast", ...)plot_pupil_activity_windows(x, feature = "pupil_frequency_contrast", ...)
x |
'eye_pupil_frequency_features' or 'eye_pupil_frequency_stability' object. |
feature |
Activity feature. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot pupil activity features across windows/groups. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot pupil low/high-band power summaries
plot_pupil_band_power(x, ...)plot_pupil_band_power(x, ...)
x |
'eye_pupil_frequency_features' object. |
... |
Additional arguments passed to the underlying method or helper. |
An R object containing plot pupil low/high-band power summaries. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Plot tonic/phasic pupil components
plot_pupil_components( data, time = "time_ms", smoothed = "pupil_smoothed", tonic = "pupil_tonic", phasic = "pupil_phasic", ... )plot_pupil_components( data, time = "time_ms", smoothed = "pupil_smoothed", tonic = "pupil_tonic", phasic = "pupil_phasic", ... )
data |
Data frame containing the required process variables. |
time |
Time values or name of the time variable. |
smoothed |
Name of the smoothed pupil-signal column. |
tonic |
Name of the tonic pupil-component column. |
phasic |
Name of the phasic pupil-component column. |
... |
Additional arguments passed to the underlying method or helper. |
A tabular R object containing plot tonic/phasic pupil components; rows represent analysis units and columns contain the returned quantities.
Plot raw-to-processed pupil preprocessing stages
plot_pupil_preprocessing_audit( data, time = "time_ms", signals = c("pupil_raw", "pupil_interpolated", "pupil_smoothed", "pupil_bc"), ... )plot_pupil_preprocessing_audit( data, time = "time_ms", signals = c("pupil_raw", "pupil_interpolated", "pupil_smoothed", "pupil_bc"), ... )
data |
Sample-level data. |
time |
Time column. |
signals |
Signal columns to overlay. |
... |
Additional arguments passed to the underlying method or helper. |
A tabular R object containing plot raw-to-processed pupil preprocessing stages; rows represent analysis units and columns contain the returned quantities.
Plot a pupil-signal power spectrum
plot_pupil_spectrum(signal, sampling_rate_hz, max_hz = sampling_rate_hz/2, ...)plot_pupil_spectrum(signal, sampling_rate_hz, max_hz = sampling_rate_hz/2, ...)
signal |
Numeric pupil signal. |
sampling_rate_hz |
Sampling rate. |
max_hz |
Maximum frequency shown; defaults to Nyquist. |
... |
Additional arguments passed to the underlying method or helper. |
A data frame containing plot a pupil-signal power spectrum. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
plot_sbc_rank(x, parameter = NULL, bins = 10L, engine = c("auto", "ggplot2", "base"), ...)plot_sbc_rank(x, parameter = NULL, bins = 10L, engine = c("auto", "ggplot2", "base"), ...)
x |
Validation collection or draw data frame. |
parameter |
Optional parameter. |
bins |
Number of bins. |
engine |
Plot engine. |
... |
Additional plotting arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
plot_validation_failures(x, engine = c("auto", "ggplot2", "base"), ...)plot_validation_failures(x, engine = c("auto", "ggplot2", "base"), ...)
x |
Validation collection or failure summary. |
engine |
Plot engine. |
... |
Additional plotting arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
plot_validation_runtime(x, engine = c("auto", "ggplot2", "base"), ...)plot_validation_runtime(x, engine = c("auto", "ggplot2", "base"), ...)
x |
Validation collection or job table. |
engine |
Plot engine. |
... |
Additional plotting arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot eye adapter regression audit
## S3 method for class 'eye_adapter_regression_audit' plot(x, ...)## S3 method for class 'eye_adapter_regression_audit' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye adapter regression audit.
Plot aoi growth curve diagnostics
## S3 method for class 'eye_aoi_growth_curve' plot(x, ...)## S3 method for class 'eye_aoi_growth_curve' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot aoi growth curve diagnostics.
Plot aoi trajectory diagnostics
## S3 method for class 'eye_aoi_trajectory' plot(x, type = c("coefficients", "profiles"), ...)## S3 method for class 'eye_aoi_trajectory' plot(x, type = c("coefficients", "profiles"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot aoi trajectory diagnostics.
Plot bayesian process dashboard diagnostics
## S3 method for class 'eye_bayesian_process_dashboard' plot(x, type = c("loo", "rhat", "ess"), ...)## S3 method for class 'eye_bayesian_process_dashboard' plot(x, type = c("loo", "rhat", "ess"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot bayesian process dashboard diagnostics.
Plot eye bids roundtrip
## S3 method for class 'eye_bids_roundtrip' plot(x, ...)## S3 method for class 'eye_bids_roundtrip' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye bids roundtrip.
Plot biometric imputation sensitivity diagnostics
## S3 method for class 'eye_biometric_imputation_sensitivity' plot(x, ...)## S3 method for class 'eye_biometric_imputation_sensitivity' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot biometric imputation sensitivity diagnostics.
Plot biometric preflight diagnostics
## S3 method for class 'eye_biometric_preflight' plot(x, type = c("heatmap", "decision_counts"), ...)## S3 method for class 'eye_biometric_preflight' plot(x, type = c("heatmap", "decision_counts"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot biometric preflight diagnostics.
Plot candidate item bank audit diagnostics
## S3 method for class 'eye_candidate_item_bank_audit' plot(x, ...)## S3 method for class 'eye_candidate_item_bank_audit' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot candidate item bank audit diagnostics.
Plot detailed compatibility evidence
## S3 method for class 'eye_compatibility_evidence_matrix' plot(x, ...)## S3 method for class 'eye_compatibility_evidence_matrix' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot detailed compatibility evidence.
Plot decision process proxy diagnostics
## S3 method for class 'eye_decision_process_proxy' plot(x, ...)## S3 method for class 'eye_decision_process_proxy' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot decision process proxy diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_dynamic_irtree' plot(x, type = c("observed", "fitted", "residual"), ...)## S3 method for class 'eye_dynamic_irtree' plot(x, type = c("observed", "fitted", "residual"), ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
type |
Value for 'type'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_dynamic_ppc' plot(x, ...)## S3 method for class 'eye_dynamic_ppc' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot eye event roundtrip audit
## S3 method for class 'eye_event_roundtrip_audit' plot(x, ...)## S3 method for class 'eye_event_roundtrip_audit' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye event roundtrip audit.
Plot eye event time irt
## S3 method for class 'eye_event_time_irt' plot(x, ...)## S3 method for class 'eye_event_time_irt' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye event time irt.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_diagnostics' plot(x, ...)## S3 method for class 'eye_functional_pupil_diagnostics' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_irt' plot(x, type = c("trajectories", "coefficients", "recovery"), ...)## S3 method for class 'eye_functional_pupil_irt' plot(x, type = c("trajectories", "coefficients", "recovery"), ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
type |
Value for 'type'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_sensitivity' plot(x, parameter = NULL, ...)## S3 method for class 'eye_functional_pupil_sensitivity' plot(x, parameter = NULL, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
parameter |
Value for 'parameter'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot gated process model diagnostics
## S3 method for class 'eye_gated_process_model' plot(x, ...)## S3 method for class 'eye_gated_process_model' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot gated process model diagnostics.
Plot gaze anchored 3pl audit diagnostics
## S3 method for class 'eye_gaze_anchored_3pl_audit' plot( x, type = c("lower_asymptote", "process_alignment", "difficulty_discrimination"), feature = NULL, ... )## S3 method for class 'eye_gaze_anchored_3pl_audit' plot( x, type = c("lower_asymptote", "process_alignment", "difficulty_discrimination"), feature = NULL, ... )
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
feature |
Process feature to evaluate or display. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot gaze anchored 3pl audit diagnostics.
Plot eye gaze informed missingness irt
## S3 method for class 'eye_gaze_informed_missingness_irt' plot(x, ...)## S3 method for class 'eye_gaze_informed_missingness_irt' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye gaze informed missingness irt.
Plot flexible IRF shape diagnostics
## S3 method for class 'eye_gpirt' plot(x, item = NULL, ...)## S3 method for class 'eye_gpirt' plot(x, item = NULL, ...)
x |
Object to print, plot, summarize, or audit. |
item |
Item identifier, name, or item column. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot flexible IRF shape diagnostics.
Plot incremental process-channel information by fold
## S3 method for class 'eye_incremental_information_audit' plot(x, ...)## S3 method for class 'eye_incremental_information_audit' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot incremental process-channel information by fold.
Plot detected process changepoints
## S3 method for class 'eye_irt_changepoints' plot(x, person = NULL, ...)## S3 method for class 'eye_irt_changepoints' plot(x, person = NULL, ...)
x |
Object to print, plot, summarize, or audit. |
person |
Person or participant identifier column. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot detected process changepoints.
Plot an IRT linking/equating transformation
## S3 method for class 'eye_irt_equating' plot(x, theta = seq(-4, 4, length.out = 201), ...)## S3 method for class 'eye_irt_equating' plot(x, theta = seq(-4, 4, length.out = 201), ...)
x |
Object to print, plot, summarize, or audit. |
theta |
Latent-trait values. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot an IRT linking/equating transformation.
Plot posterior predictive discrepancy tail probabilities
## S3 method for class 'eye_irt_ppc' plot(x, ...)## S3 method for class 'eye_irt_ppc' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot posterior predictive discrepancy tail probabilities.
Plot parameter-recovery bias or RMSE
## S3 method for class 'eye_irt_recovery_summary' plot(x, metric = c("rmse", "absolute_bias", "coverage"), ...)## S3 method for class 'eye_irt_recovery_summary' plot(x, metric = c("rmse", "absolute_bias", "coverage"), ...)
x |
Object to print, plot, summarize, or audit. |
metric |
Metric to calculate or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot parameter-recovery bias or RMSE.
Plot SBC rank histograms by parameter
## S3 method for class 'eye_irt_sbc' plot(x, parameter = NULL, breaks = 10L, ...)## S3 method for class 'eye_irt_sbc' plot(x, parameter = NULL, breaks = 10L, ...)
x |
Object to print, plot, summarize, or audit. |
parameter |
Value supplied to 'parameter'; see Details for its model-specific role. |
breaks |
Histogram or discretization breaks. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot SBC rank histograms by parameter.
Plot item parameter seed diagnostics
## S3 method for class 'eye_item_parameter_seed' plot(x, candidate_data = NULL, ...)## S3 method for class 'eye_item_parameter_seed' plot(x, candidate_data = NULL, ...)
x |
Object to process, inspect, compare, or plot. |
candidate_data |
Optional candidate-item data used for the requested display. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot item parameter seed diagnostics.
Plot item reduction sensitivity diagnostics
## S3 method for class 'eye_item_reduction_sensitivity' plot(x, ...)## S3 method for class 'eye_item_reduction_sensitivity' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot item reduction sensitivity diagnostics.
Plot a joint gaze-response-time IRT fit
## S3 method for class 'eye_joint_gaze_rt_irt' plot(x, type = c("latent", "components"), ...)## S3 method for class 'eye_joint_gaze_rt_irt' plot(x, type = c("latent", "components"), ...)
x |
Object to print, plot, summarize, or audit. |
type |
Plot or result type. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot a joint gaze-response-time IRT fit.
Plot a graded response + RT/process fit
## S3 method for class 'eye_joint_graded_rt_process_irt' plot(x, ...)## S3 method for class 'eye_joint_graded_rt_process_irt' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot a graded response + RT/process fit.
Plot eye latent distribution comparison
## S3 method for class 'eye_latent_distribution_comparison' plot(x, ...)## S3 method for class 'eye_latent_distribution_comparison' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye latent distribution comparison.
Plot latent process alignment diagnostics
## S3 method for class 'eye_latent_process_alignment' plot(x, ...)## S3 method for class 'eye_latent_process_alignment' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot latent process alignment diagnostics.
Plot a latent-space IRT adapter fit
## S3 method for class 'eye_latent_space_irt' plot(x, ...)## S3 method for class 'eye_latent_space_irt' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot a latent-space IRT adapter fit.
Plot many-facet process IRT effects
## S3 method for class 'eye_manyfacet_process_irt' plot(x, facet = NULL, ...)## S3 method for class 'eye_manyfacet_process_irt' plot(x, facet = NULL, ...)
x |
Object to print, plot, summarize, or audit. |
facet |
Facet to display or extract. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot many-facet process IRT effects.
Plot mixture irt process diagnostics
## S3 method for class 'eye_mixture_irt_process' plot(x, ...)## S3 method for class 'eye_mixture_irt_process' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot mixture irt process diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_model_promotion_audit' plot(x, ...)## S3 method for class 'eye_model_promotion_audit' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot multiblock process map diagnostics
## S3 method for class 'eye_multiblock_process_map' plot(x, type = c("individuals", "variables", "blocks", "contributions"), ...)## S3 method for class 'eye_multiblock_process_map' plot(x, type = c("individuals", "variables", "blocks", "contributions"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot multiblock process map diagnostics.
Plot nominal-response gaze results
## S3 method for class 'eye_nominal_gaze_irt' plot(x, type = c("distractor_map", "coefficients"), ...)## S3 method for class 'eye_nominal_gaze_irt' plot(x, type = c("distractor_map", "coefficients"), ...)
x |
Object to print, plot, summarize, or audit. |
type |
Plot or result type. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot nominal-response gaze results.
Plot nonparametric rasch audit diagnostics
## S3 method for class 'eye_nonparametric_rasch_audit' plot(x, method = names(x$tests)[1L], ...)## S3 method for class 'eye_nonparametric_rasch_audit' plot(x, method = names(x$tests)[1L], ...)
x |
Object to process, inspect, compare, or plot. |
method |
Method used for the requested diagnostic or summary. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot nonparametric rasch audit diagnostics.
Plot omission/not-reached survival IRT diagnostics
## S3 method for class 'eye_omission_survival_irt' plot(x, type = c("survival", "missingness"), ...)## S3 method for class 'eye_omission_survival_irt' plot(x, type = c("survival", "missingness"), ...)
x |
Object to print, plot, summarize, or audit. |
type |
Plot or result type. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot omission/not-reached survival IRT diagnostics.
Plot preaction process features diagnostics
## S3 method for class 'eye_preaction_process_features' plot(x, feature = "pupil_mean", ...)## S3 method for class 'eye_preaction_process_features' plot(x, feature = "pupil_mean", ...)
x |
Object to process, inspect, compare, or plot. |
feature |
Process feature to evaluate or display. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot preaction process features diagnostics.
Plot presentation accessibility diagnostics
## S3 method for class 'eye_presentation_accessibility' plot(x, ...)## S3 method for class 'eye_presentation_accessibility' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot presentation accessibility diagnostics.
Plot presentation fairness comparison diagnostics
## S3 method for class 'eye_presentation_fairness_comparison' plot(x, ...)## S3 method for class 'eye_presentation_fairness_comparison' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot presentation fairness comparison diagnostics.
Plot process anomaly audit diagnostics
## S3 method for class 'eye_process_anomaly_audit' plot(x, ...)## S3 method for class 'eye_process_anomaly_audit' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process anomaly audit diagnostics.
Plot CAT simulation information accumulation
## S3 method for class 'eye_process_cat_simulation' plot(x, ...)## S3 method for class 'eye_process_cat_simulation' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot CAT simulation information accumulation.
Plot process-channel ablation
## S3 method for class 'eye_process_channel_ablation' plot(x, ...)## S3 method for class 'eye_process_channel_ablation' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process-channel ablation.
Plot process-dependent discrimination
## S3 method for class 'eye_process_dependent_discrimination' plot(x, ...)## S3 method for class 'eye_process_dependent_discrimination' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process-dependent discrimination.
Plot process drift audit diagnostics
## S3 method for class 'eye_process_drift_audit' plot( x, type = c("trajectory", "delta", "heatmap", "control"), metric = NULL, item = NULL, ... )## S3 method for class 'eye_process_drift_audit' plot( x, type = c("trajectory", "delta", "heatmap", "control"), metric = NULL, item = NULL, ... )
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
metric |
Metric to evaluate or display. |
item |
Optional item identifier used to restrict or highlight results. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process drift audit diagnostics.
Plot process external validity diagnostics
## S3 method for class 'eye_process_external_validity' plot( x, type = c("associations", "incremental", "observed_fitted", "residuals"), ... )## S3 method for class 'eye_process_external_validity' plot( x, type = c("associations", "incremental", "observed_fitted", "residuals"), ... )
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process external validity diagnostics.
Plot eye process facet effects
## S3 method for class 'eye_process_facet_effects' plot(x, ...)## S3 method for class 'eye_process_facet_effects' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye process facet effects.
Plot process feature blocks diagnostics
## S3 method for class 'eye_process_feature_blocks' plot(x, ...)## S3 method for class 'eye_process_feature_blocks' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process feature blocks diagnostics.
Plot process-measure variance components
## S3 method for class 'eye_process_g_study' plot(x, ...)## S3 method for class 'eye_process_g_study' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process-measure variance components.
Plot a process-HMM IRT fit
## S3 method for class 'eye_process_hmm_irt' plot(x, type = c("occupancy", "transition"), ...)## S3 method for class 'eye_process_hmm_irt' plot(x, type = c("occupancy", "transition"), ...)
x |
Object to print, plot, summarize, or audit. |
type |
Plot or result type. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot a process-HMM IRT fit.
Plot process/local-dependence diagnostics
## S3 method for class 'eye_process_local_dependence_audit' plot(x, ...)## S3 method for class 'eye_process_local_dependence_audit' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process/local-dependence diagnostics.
Plot a process-channel negative-control distribution
## S3 method for class 'eye_process_negative_control' plot(x, ...)## S3 method for class 'eye_process_negative_control' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot a process-channel negative-control distribution.
Plot process person-fit discrepancies
## S3 method for class 'eye_process_person_fit' plot(x, top = 25L, ...)## S3 method for class 'eye_process_person_fit' plot(x, top = 25L, ...)
x |
Object to print, plot, summarize, or audit. |
top |
Number of highest-ranked cases to display. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process person-fit discrepancies.
Plot process profile mixture diagnostics
## S3 method for class 'eye_process_profile_mixture' plot(x, type = c("profiles", "posterior", "scatter", "parallel"), ...)## S3 method for class 'eye_process_profile_mixture' plot(x, type = c("profiles", "posterior", "scatter", "parallel"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process profile mixture diagnostics.
Plot process rasch tree diagnostics
## S3 method for class 'eye_process_rasch_tree' plot(x, ...)## S3 method for class 'eye_process_rasch_tree' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process rasch tree diagnostics.
Plot process window sensitivity diagnostics
## S3 method for class 'eye_process_window_sensitivity' plot(x, ...)## S3 method for class 'eye_process_window_sensitivity' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process window sensitivity diagnostics.
Plot process windows diagnostics
## S3 method for class 'eye_process_windows' plot(x, feature = "pupil_mean", group = NULL, ...)## S3 method for class 'eye_process_windows' plot(x, feature = "pupil_mean", group = NULL, ...)
x |
Object to process, inspect, compare, or plot. |
feature |
Process feature to evaluate or display. |
group |
Optional grouping variable. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot process windows diagnostics.
Plot pupil confound model diagnostics
## S3 method for class 'eye_pupil_confound_model' plot( x, type = c("luminance", "trial_order", "raw_adjusted", "theta_luminance_surface"), ... )## S3 method for class 'eye_pupil_confound_model' plot( x, type = c("luminance", "trial_order", "raw_adjusted", "theta_luminance_surface"), ... )
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot pupil confound model diagnostics.
Plot pupil deconvolution diagnostics
## S3 method for class 'eye_pupil_deconvolution' plot(x, type = c("observed_fitted", "effects", "residuals", "kernels"), ...)## S3 method for class 'eye_pupil_deconvolution' plot(x, type = c("observed_fitted", "effects", "residuals", "kernels"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot pupil deconvolution diagnostics.
Plot pupil fatigue drift diagnostics
## S3 method for class 'eye_pupil_fatigue_drift' plot(x, ...)## S3 method for class 'eye_pupil_fatigue_drift' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot pupil fatigue drift diagnostics.
Plot pupil frequency features diagnostics
## S3 method for class 'eye_pupil_frequency_features' plot(x, type = c("features", "power_relationship"), ...)## S3 method for class 'eye_pupil_frequency_features' plot(x, type = c("features", "power_relationship"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot pupil frequency features diagnostics.
Plot pupil frequency stability diagnostics
## S3 method for class 'eye_pupil_frequency_stability' plot(x, feature = "pupil_frequency_contrast", ...)## S3 method for class 'eye_pupil_frequency_stability' plot(x, feature = "pupil_frequency_contrast", ...)
x |
Object to process, inspect, compare, or plot. |
feature |
Process feature to evaluate or display. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot pupil frequency stability diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_roundtrip_loss_audit' plot(x, ...)## S3 method for class 'eye_roundtrip_loss_audit' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot SBC audit summaries
## S3 method for class 'eye_sbc_audit' plot(x, ...)## S3 method for class 'eye_sbc_audit' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot SBC audit summaries.
Plot semantic round-trip fidelity
## S3 method for class 'eye_semantic_roundtrip' plot(x, ...)## S3 method for class 'eye_semantic_roundtrip' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot semantic round-trip fidelity.
Plot signal filter audit diagnostics
## S3 method for class 'eye_signal_filter_audit' plot(x, ...)## S3 method for class 'eye_signal_filter_audit' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot signal filter audit diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_storage_benchmark' plot(x, metric = c("write_seconds", "read_seconds", "bytes"), ...)## S3 method for class 'eye_storage_benchmark' plot(x, metric = c("write_seconds", "read_seconds", "bytes"), ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
metric |
Value for 'metric'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_strategy_aoi_sensitivity' plot(x, ...)## S3 method for class 'eye_strategy_aoi_sensitivity' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot streaming score diagnostics
## S3 method for class 'eye_streaming_score' plot(x, ...)## S3 method for class 'eye_streaming_score' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot streaming score diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_transition_diagnostics' plot(x, ...)## S3 method for class 'eye_transition_diagnostics' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot validation bundle diagnostics
## S3 method for class 'eye_validation_bundle' plot(x, ...)## S3 method for class 'eye_validation_bundle' plot(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot validation bundle diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_vendor_compatibility_matrix' plot(x, ...)## S3 method for class 'eye_vendor_compatibility_matrix' plot(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Plot eye vendor semantic validation
## S3 method for class 'eye_vendor_semantic_validation' plot(x, ...)## S3 method for class 'eye_vendor_semantic_validation' plot(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the plotting result when available; the primary effect is drawing plot eye vendor semantic validation.
Plot visual context irt diagnostics
## S3 method for class 'eye_visual_context_irt' plot(x, type = c("loadings", "difficulty_change", "context_registry"), ...)## S3 method for class 'eye_visual_context_irt' plot(x, type = c("loadings", "difficulty_change", "context_registry"), ...)
x |
Object to process, inspect, compare, or plot. |
type |
Type of summary or visual representation to produce. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the plotting result when available; the primary effect is drawing plot visual context irt diagnostics.
Posterior predictive discrepancy table
posterior_predictive_discrepancies( observed, replicated, discrepancies = list(mean = function(x) mean(x, na.rm = TRUE), sd = function(x) stats::sd(x, na.rm = TRUE), zero_rate = function(x) mean(x == 0, na.rm = TRUE)) )posterior_predictive_discrepancies( observed, replicated, discrepancies = list(mean = function(x) mean(x, na.rm = TRUE), sd = function(x) stats::sd(x, na.rm = TRUE), zero_rate = function(x) mean(x == 0, na.rm = TRUE)) )
observed |
Observed vector/data object. |
replicated |
List of replicated datasets, or matrix with one replicate per row. |
discrepancies |
Named list of functions mapping a dataset to one number. |
An object of class "eye_irt_ppc", "data.frame", stored as a data frame, containing posterior predictive discrepancy table and associated metadata needed to interpret the result.
Posterior SBC is deliberately callback-driven: conditioning/augmentation differs by model, and eyeprocess does not pretend that re-running ordinary prior SBC on an observed-data neighbourhood is posterior SBC.
posterior_sbc_contract(replication)posterior_sbc_contract(replication)
replication |
Function '(replicate, observed_data)' returning 'list(truth=named_numeric, draws=matrix_or_data_frame)'. The callback is responsible for the conditional posterior-SBC construction appropriate to the model, including fitting to the observed data and the required self-consistency experiment. |
An object of class "eye_posterior_sbc_contract", stored as a named list, with components "replication", "requirement". It contains define a posterior-SBC replication contract and associated metadata or diagnostics needed to interpret the result.
Build pre-action process features
preaction_process_features( data, by = c("person_id", "trial_id"), time = "time_ms", response_time = "response_time_ms", windows_ms = c(500, 1000, 2000), aoi = "aoi", pupil = "pupil_bc", blink = "blink" )preaction_process_features( data, by = c("person_id", "trial_id"), time = "time_ms", response_time = "response_time_ms", windows_ms = c(500, 1000, 2000), aoi = "aoi", pupil = "pupil_bc", blink = "blink" )
data |
Sample-level data. |
by |
Grouping columns. |
time, response_time
|
Time and response-event columns. |
windows_ms |
Positive look-back windows before action. |
aoi, pupil, blink
|
Optional feature columns. |
An object of class "eye_preaction_process_features", stored as a named list, with components "data", "windows_ms", "by", "status", "caveat". It contains pre-action process features and associated metadata or diagnostics needed to interpret the result.
Predict from an AOI growth curve
predict_aoi_trajectory(object, time = NULL)predict_aoi_trajectory(object, time = NULL)
object |
Object supplied to the S3 method. |
time |
Time values or name of the time variable. |
A data frame containing from an AOI growth curve. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Predict pre-pilot item-parameter priors
predict_item_parameter_priors(object, newdata)predict_item_parameter_priors(object, newdata)
object |
Seed model. |
newdata |
Candidate item feature data. |
An R object containing pre-pilot item-parameter priors. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Predict a latent trait at arbitrary times
predict_theta_at_time(object, time)predict_theta_at_time(object, time)
object |
A fitted eyeprocess model or audit object. |
time |
Time values. |
An R object containing a latent trait at arbitrary times. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Predict expected bounded response from a censored-normal process IRT fit
## S3 method for class 'eye_censored_normal_process_irt' predict(object, theta = object$theta, items = NULL, ...)## S3 method for class 'eye_censored_normal_process_irt' predict(object, theta = object$theta, items = NULL, ...)
object |
A fitted eyeprocess model or audit object. |
theta |
Latent-trait values. |
items |
Items to include. |
... |
Additional arguments passed to the selected model, engine, or method. |
A vector or matrix containing expected bounded response from a censored-normal process IRT fit, with shape determined by the supplied analysis units.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_multinomial_transition' predict(object, newdata = NULL, type = c("probability", "class", "link"), ...)## S3 method for class 'eye_multinomial_transition' predict(object, newdata = NULL, type = c("probability", "class", "link"), ...)
object |
Multinomial transition model. |
newdata |
Optional prepared transition data. |
type |
Probability, class, or linear predictor. |
... |
Unused. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Extract pre-flight decisions
preflight_decisions(x)preflight_decisions(x)
x |
An 'eye_biometric_preflight' object. |
An R object containing pre-flight decisions. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
The manifest records recommendations only. It does not remove observations.
preflight_exclusion_manifest(x)preflight_exclusion_manifest(x)
x |
Object to process, inspect, compare, or plot. |
A tabular R object containing an explicit pre-flight exclusion/review manifest; rows represent analysis units and columns contain the returned quantities.
Extract pre-flight failures/review cases
preflight_failures(x)preflight_failures(x)
x |
Object to process, inspect, compare, or plot. |
A tabular R object containing pre-flight failures/review cases; rows represent analysis units and columns contain the returned quantities.
Extract pre-flight passes
preflight_passed(x)preflight_passed(x)
x |
Object to process, inspect, compare, or plot. |
A tabular R object containing pre-flight passes; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
prepare_dynamic_irtree_data(x, spec = dynamic_irtree_spec(), person = "participant_id", item = "item_id", trial = "trial_id", state = "state", time = NULL, from = "from_state", to = "to_state", states = NULL)prepare_dynamic_irtree_data(x, spec = dynamic_irtree_spec(), person = "participant_id", item = "item_id", trial = "trial_id", state = "state", time = NULL, from = "from_state", to = "to_state", states = NULL)
x |
An 'eye_dataset' or data frame. |
spec |
Dynamic IRTree specification. |
person |
Column names for long state data. |
item |
Column names for long state data. |
trial |
Column names for long state data. |
state |
Column names for long state data. |
time |
Column names for long state data. |
from |
Existing transition columns for transition-format data. |
to |
Existing transition columns for transition-format data. |
states |
Optional complete state vocabulary. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
prepare_functional_pupil_data(x, spec = functional_pupil_irt_spec())prepare_functional_pupil_data(x, spec = functional_pupil_irt_spec())
x |
Eye dataset or long pupil data. |
spec |
Functional pupil specification. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
prepare_gaze_diffusion_data(data, spec, minimum_rt = 0.05)prepare_gaze_diffusion_data(data, spec, minimum_rt = 0.05)
data |
Trial-level data. |
spec |
Diffusion specification. |
minimum_rt |
Minimum admissible RT in seconds. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Creates a loss-aware person-item-trial table for response, response time, gaze, and pupil channels while retaining references to optional sample-level or sequence payloads. No time series are silently aggregated.
prepare_multimodal_irt_data( data, person, item, trial = NULL, response = NULL, rt = NULL, gaze = NULL, pupil = NULL, quality = character(), device = NULL, payloads = list(), provenance = list() )prepare_multimodal_irt_data( data, person, item, trial = NULL, response = NULL, rt = NULL, gaze = NULL, pupil = NULL, quality = character(), device = NULL, payloads = list(), provenance = list() )
data |
A data.frame containing one row per intended person-item-trial. |
person, item, trial
|
Column names identifying the measurement keys. |
response, rt, gaze, pupil
|
Optional column names for channel summaries. |
quality |
Optional character vector of quality-field names. |
device |
Optional list or data.frame of device metadata. |
payloads |
Named list of sample-level/sequence payloads. |
provenance |
Optional provenance list. |
An 'eye_multimodal_measurement' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
prepare_strategy_mixture_data(data, spec, standardize = TRUE)prepare_strategy_mixture_data(data, spec, standardize = TRUE)
data |
Trial-level data. |
spec |
Strategy specification. |
standardize |
Whether to standardize process features. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Prepare leakage-safe structured/unstructured process representations
prepare_structured_unstructured_process_features( structured, unstructured = NULL, fold = NULL, builder = NULL, id = c("person_id", "item_id"), ... )prepare_structured_unstructured_process_features( structured, unstructured = NULL, fold = NULL, builder = NULL, id = c("person_id", "item_id"), ... )
structured |
Structured person/item/window features. |
unstructured |
Optional sequence/sample object. |
fold |
Fold identifier. If supplied, representation builders are applied fold-locally using 'builder'. |
builder |
Optional function '(train_structured, train_unstructured, test_structured, test_unstructured, fold_value, ...)' returning a fold result. |
id |
Optional identifier columns retained in the representation contract. |
... |
Passed to 'builder'. |
An object of class "eye_structured_unstructured_process_features", stored as a named list, with components "contract", "folds", "status". It contains leakage-safe structured/unstructured process representations and associated metadata or diagnostics needed to interpret the result.
Run a preprocessing or AOI multiverse
preprocessing_multiverse( x, specifications, transform, analyse, extract = function(z) as.data.frame(z) )preprocessing_multiverse( x, specifications, transform, analyse, extract = function(z) as.data.frame(z) )
x |
Input object. |
specifications |
Named list of specification objects. |
transform |
Function receiving 'x' and one specification. |
analyse |
Analysis function receiving the transformed object. |
extract |
Function converting an analysis result to a data frame. |
An 'eye_multiverse' object.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_benchmark_reproduction' print(x, ...)## S3 method for class 'eye_benchmark_reproduction' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_benchmark_study' print(x, ...)## S3 method for class 'eye_benchmark_study' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_benchmark_validation' print(x, ...)## S3 method for class 'eye_benchmark_validation' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_diffusion_diagnostics' print(x, ...)## S3 method for class 'eye_diffusion_diagnostics' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_diffusion_identification_study' print(x, ...)## S3 method for class 'eye_diffusion_identification_study' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_dynamic_irtree' print(x, ...)## S3 method for class 'eye_dynamic_irtree' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_dynamic_ppc' print(x, ...)## S3 method for class 'eye_dynamic_ppc' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_dynamic_recovery' print(x, ...)## S3 method for class 'eye_dynamic_recovery' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_engine_adapter_result' print(x, ...)## S3 method for class 'eye_engine_adapter_result' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_data' print(x, ...)## S3 method for class 'eye_functional_pupil_data' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_diagnostics' print(x, ...)## S3 method for class 'eye_functional_pupil_diagnostics' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_irt' print(x, ...)## S3 method for class 'eye_functional_pupil_irt' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_pupil_sensitivity' print(x, ...)## S3 method for class 'eye_functional_pupil_sensitivity' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_functional_scalar_comparison' print(x, ...)## S3 method for class 'eye_functional_scalar_comparison' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print a gated process model object
## S3 method for class 'eye_gated_process_model' print(x, ...)## S3 method for class 'eye_gated_process_model' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_gaze_diffusion_data' print(x, ...)## S3 method for class 'eye_gaze_diffusion_data' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_gaze_diffusion_irt' print(x, ...)## S3 method for class 'eye_gaze_diffusion_irt' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_gaze_diffusion_spec' print(x, ...)## S3 method for class 'eye_gaze_diffusion_spec' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print eye irt evidence grade
## S3 method for class 'eye_irt_evidence_grade' print(x, ...)## S3 method for class 'eye_irt_evidence_grade' print(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Print a multimodal IRT model specification
## S3 method for class 'eye_irt_model_spec' print(x, ...)## S3 method for class 'eye_irt_model_spec' print(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Print eye irt validation spec
## S3 method for class 'eye_irt_validation_spec' print(x, ...)## S3 method for class 'eye_irt_validation_spec' print(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_model_contract_validation' print(x, ...)## S3 method for class 'eye_model_contract_validation' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_model_promotion_audit' print(x, ...)## S3 method for class 'eye_model_promotion_audit' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_multinomial_transition' print(x, ...)## S3 method for class 'eye_multinomial_transition' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_partition_spec' print(x, ...)## S3 method for class 'eye_partition_spec' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_partitioned_storage' print(x, ...)## S3 method for class 'eye_partitioned_storage' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print a process drift spec object
## S3 method for class 'eye_process_drift_spec' print(x, ...)## S3 method for class 'eye_process_drift_spec' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Print eye process negative control
## S3 method for class 'eye_process_negative_control' print(x, ...)## S3 method for class 'eye_process_negative_control' print(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Print a process preflight spec object
## S3 method for class 'eye_process_preflight_spec' print(x, ...)## S3 method for class 'eye_process_preflight_spec' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Print a process window spec object
## S3 method for class 'eye_process_window_spec' print(x, ...)## S3 method for class 'eye_process_window_spec' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_redaction_result' print(x, ...)## S3 method for class 'eye_redaction_result' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_roundtrip_loss_audit' print(x, ...)## S3 method for class 'eye_roundtrip_loss_audit' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_strategy_data' print(x, ...)## S3 method for class 'eye_strategy_data' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_strategy_manipulation_validation' print(x, ...)## S3 method for class 'eye_strategy_manipulation_validation' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_theory_strategy_irt' print(x, ...)## S3 method for class 'eye_theory_strategy_irt' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_theory_strategy_spec' print(x, ...)## S3 method for class 'eye_theory_strategy_spec' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_transition_design' print(x, ...)## S3 method for class 'eye_transition_design' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_transition_diagnostics' print(x, ...)## S3 method for class 'eye_transition_diagnostics' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print a validation bundle object
## S3 method for class 'eye_validation_bundle' print(x, ...)## S3 method for class 'eye_validation_bundle' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_case_fingerprint' print(x, ...)## S3 method for class 'eye_validation_case_fingerprint' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_collection' print(x, ...)## S3 method for class 'eye_validation_collection' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_completion_audit' print(x, ...)## S3 method for class 'eye_validation_completion_audit' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_job_plan' print(x, ...)## S3 method for class 'eye_validation_job_plan' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_run' print(x, ...)## S3 method for class 'eye_validation_run' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_vendor_case' print(x, ...)## S3 method for class 'eye_vendor_case' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_vendor_compatibility_matrix' print(x, ...)## S3 method for class 'eye_vendor_compatibility_matrix' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print eye vendor schema contract
## S3 method for class 'eye_vendor_schema_contract' print(x, ...)## S3 method for class 'eye_vendor_schema_contract' print(x, ...)
x |
Object to print, plot, summarize, or audit. |
... |
Additional arguments passed to the selected model, engine, or method. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_vendor_semantic_comparison' print(x, ...)## S3 method for class 'eye_vendor_semantic_comparison' print(x, ...)
x |
Value for 'x'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Print a visual context registry object
## S3 method for class 'eye_visual_context_registry' print(x, ...)## S3 method for class 'eye_visual_context_registry' print(x, ...)
x |
Object to process, inspect, compare, or plot. |
... |
Additional arguments passed to the underlying method or helper. |
Invisibly returns the input object after printing its summary; the object's class and contents are unchanged.
Probabilistic AOI assignment under empirical calibration uncertainty
probabilistic_aoi_assignment( data, aois, model, x = "gaze_x", y = "gaze_y", draws = 500L, seed = 1L, min_probability = 0.5 )probabilistic_aoi_assignment( data, aois, model, x = "gaze_x", y = "gaze_y", draws = 500L, seed = 1L, min_probability = 0.5 )
data |
Gaze samples. |
aois |
Rectangular AOI table. |
model |
Calibration error model. |
x, y
|
Gaze-coordinate columns. |
draws |
Monte Carlo draws. |
seed |
Seed. |
min_probability |
Minimum probability for assignment; lower maxima become 'NA'. |
An object of class "eye_probabilistic_aoi_assignment", stored as a named list, with components "assignments", "probabilities", "aois", "model", "min_probability", "caveat". It contains probabilistic AOI assignment under empirical calibration uncertainty and associated metadata or diagnostics needed to interpret the result.
Extract multivariate process anomaly distances
process_anomaly_distance(x)process_anomaly_distance(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing multivariate process anomaly distances. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Bland-Altman repeatability summary for two sessions
process_bland_altman(data, person, session, measure, sessions = NULL)process_bland_altman(data, person, session, measure, sessions = NULL)
data |
Long data. |
person |
Participant column. |
session |
Session column containing exactly two selected sessions. |
measure |
Measure column. |
sessions |
Optional two session labels. |
An object of class "eye_process_bland_altman", stored as a named list, with components "pairs", "summary", "sessions". It contains bland-Altman repeatability summary for two sessions and associated metadata or diagnostics needed to interpret the result.
Ablate process channels under a common out-of-sample evaluator
process_channel_ablation( data, channels, evaluator, baseline = character(), higher_is_better = TRUE )process_channel_ablation( data, channels, evaluator, baseline = character(), higher_is_better = TRUE )
data |
Input data. |
channels |
Named list whose elements are character vectors of columns. |
evaluator |
Function '(data, active_columns, channel_name)' returning a scalar out-of-sample score. The evaluator owns all fitting/splitting logic. |
baseline |
Character vector of always-active columns. |
higher_is_better |
Direction of the score. |
An object of class "eye_process_channel_ablation", "data.frame", stored as a data frame, containing ablate process channels under a common out-of-sample evaluator and associated metadata needed to interpret the result.
Extract process-criterion associations
process_criterion_associations(x)process_criterion_associations(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing process-criterion associations. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Residualises a process variable against person/item baselines, then tests whether the theta-response slope changes with that residual process value. This is a transparent diagnostic inspired by 2026 evidence on conditional response-time/discrimination dependencies, not an exact reproduction of the published meta-analytic model.
process_dependent_discrimination_audit( data, response, theta, process, person, item, nonlinear = TRUE )process_dependent_discrimination_audit( data, response, theta, process, person, item, nonlinear = TRUE )
data |
Long-format response data. |
response |
Binary response column. |
theta |
Person latent-score column. |
process |
RT/gaze/process measure. |
person, item
|
Identifier columns. |
nonlinear |
If TRUE and mgcv is installed, additionally estimate a smooth theta-by-process diagnostic surface. |
An object of class "eye_process_dependent_discrimination", stored as a named list, with components "process_model", "response_model", "interaction", "smooth_model", "residual_process", "status", "caveat". It contains process-dependent item discrimination and associated metadata or diagnostics needed to interpret the result.
Construct a process-data nuisance surrogate for DIF analysis
process_dif_nuisance_surrogate( data, process_features, person = "participant_id", aggregate = TRUE )process_dif_nuisance_surrogate( data, process_features, person = "participant_id", aggregate = TRUE )
data |
Input data frame or compatible tabular object. |
process_features |
Names of process-derived features. |
person |
Person or participant identifier column. |
aggregate |
Aggregation rule for process features. |
A data frame containing a process-data nuisance surrogate for DIF analysis. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Extract drift alerts
process_drift_alerts(x)process_drift_alerts(x)
x |
Object to process, inspect, compare, or plot. |
A logical value or vector indicating drift alerts.
Specify a process-deployment drift audit
process_drift_spec( baseline = c("first_batch", "reference_batch"), difficulty_limit = 0.4, discrimination_limit = 0.35, gaze_validity_drop = 0.1, luminance_limit = 25, relative_metric_quantile = 0.9, min_batches = 2L )process_drift_spec( baseline = c("first_batch", "reference_batch"), difficulty_limit = 0.4, discrimination_limit = 0.35, gaze_validity_drop = 0.1, luminance_limit = 25, relative_metric_quantile = 0.9, min_batches = 2L )
baseline |
Baseline rule: first observed batch or an explicitly supplied reference batch. |
difficulty_limit |
Absolute item-difficulty change triggering review. |
discrimination_limit |
Absolute discrimination change triggering review. |
gaze_validity_drop |
Maximum tolerated decrease in gaze validity. |
luminance_limit |
Absolute luminance change triggering review. |
relative_metric_quantile |
Quantile of absolute deltas used for process metrics without an externally meaningful absolute threshold. |
min_batches |
Minimum number of batches per item for drift assessment. |
An 'eye_process_drift_spec' object.
Define conceptual process-feature blocks
process_feature_blocks(data, blocks, id = NULL, drop_constant = TRUE)process_feature_blocks(data, blocks, id = NULL, drop_constant = TRUE)
data |
Data frame. |
blocks |
Named list of column names for conceptual blocks. |
id |
Optional identifier column. |
drop_constant |
Remove non-varying columns. |
An object of class "eye_process_feature_blocks", stored as a named list, with components "data", "blocks", "id", "block_sizes", "status". It contains define conceptual process-feature blocks and associated metadata or diagnostics needed to interpret the result.
Registry of process-feature families and interpretation guardrails
process_feature_family_registry()process_feature_family_registry()
A data frame containing registry of process-feature families and interpretation guardrails. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Summarize process-feature stability across repeated analyses
process_feature_stability( data, feature = "feature", split = "split", importance = "importance", top_n = 20L )process_feature_stability( data, feature = "feature", split = "split", importance = "importance", top_n = 20L )
data |
Long table containing feature names and ranks/importance values. |
feature |
Feature column. |
split |
Split/resample column. |
importance |
Importance column, where larger is better. |
top_n |
Number of top features counted per split. |
A data frame containing process-feature stability across repeated analyses. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Declare temporal provenance for process features
process_feature_time_provenance( feature, available_at, outcome_at, source = NA_character_, transformation = NA_character_, unit = "ms" )process_feature_time_provenance( feature, available_at, outcome_at, source = NA_character_, transformation = NA_character_, unit = "ms" )
feature |
Feature names. |
available_at |
Earliest time at which each feature is available. |
outcome_at |
Time at which the modeled outcome becomes available. |
source |
Optional source labels. |
transformation |
Optional transformation descriptions. |
unit |
Time unit label. |
A provenance table.
Absolute-agreement ICC(A,1) for repeated process measures
process_icc(data, person, session, measure)process_icc(data, person, session, measure)
data |
Long data. |
person |
Participant column. |
session |
Session/repetition column. |
measure |
Measure column. |
A data frame containing absolute-agreement ICC(A,1) for repeated process measures. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compares posterior draws for the same latent target under a baseline and an augmented model. This implementation intentionally uses posterior uncertainty metrics rather than assuming Fisher information is additively decomposable across heterogeneous channels.
process_information( baseline, augmented, metric = c("variance_reduction", "precision_gain", "entropy_reduction") )process_information( baseline, augmented, metric = c("variance_reduction", "precision_gain", "entropy_reduction") )
baseline, augmented
|
Numeric posterior draws. Rows are draws and columns are matched latent targets. |
metric |
'variance_reduction', 'precision_gain', or 'entropy_reduction'. |
An 'eye_process_information' data.frame.
2PL response item information
process_item_information( theta, a, b, process_information = 0, rt_information = 0, weights = c(response = 1, rt = 0, process = 0), expected_time = 0, burden_weight = 0 )process_item_information( theta, a, b, process_information = 0, rt_information = 0, weights = c(response = 1, rt = 0, process = 0), expected_time = 0, burden_weight = 0 )
theta |
Latent-trait values. |
a |
Item discrimination parameter or parameters. |
b |
Item difficulty/location parameter or parameters. |
process_information |
Information supplied by the process channel. |
rt_information |
Information supplied by response time. |
weights |
Weights used to combine information components. |
expected_time |
Expected response time or burden. |
burden_weight |
Penalty applied to expected burden. |
An object of class "eye_process_item_information", stored as a named list, with components "theta", "response_information", "utility". It contains 2PL response item information and associated metadata or diagnostics needed to interpret the result.
Return a one-measure process card
process_measure_card(name, registry = process_measure_registry())process_measure_card(name, registry = process_measure_registry())
name |
Measure name. |
registry |
Registry. |
An object of class "eye_process_measure_card", stored as a named list, containing return a one-measure process card and associated metadata needed to interpret the result.
Process-measure coverage for an observed dataset
process_measure_coverage(data, registry = process_measure_registry())process_measure_coverage(data, registry = process_measure_registry())
data |
Data frame. |
registry |
Registry. |
A data frame containing process-measure coverage for an observed dataset. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Process-measure guardrail table
process_measure_guardrails(registry = process_measure_registry())process_measure_guardrails(registry = process_measure_registry())
registry |
Registry. |
An R object containing process-measure guardrail table. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Process-measure lineage table
process_measure_lineage( measure, inputs, transformations = character(), output_level = NA_character_ )process_measure_lineage( measure, inputs, transformations = character(), output_level = NA_character_ )
measure |
Measure name. |
inputs |
Required input variable names. |
transformations |
Ordered transformation labels. |
output_level |
Output aggregation level. |
An object of class "eye_process_measure_lineage", stored as a named list, with components "measure", "inputs", "transformations", "output_level", "lineage_hash". It contains process-measure lineage table and associated metadata or diagnostics needed to interpret the result.
Unified process-measure registry
process_measure_registry(include_experimental = TRUE)process_measure_registry(include_experimental = TRUE)
include_experimental |
Include experimental registry entries. |
A data frame containing unified process-measure registry. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
List units used by registered process measures
process_measure_units(registry = process_measure_registry())process_measure_units(registry = process_measure_registry())
registry |
Registry. |
An R object containing list units used by registered process measures. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Permutation negative control
process_negative_control_permute(data, outcome, seed = 1L, within = NULL)process_negative_control_permute(data, outcome, seed = 1L, within = NULL)
data |
Data frame. |
outcome |
Outcome column. |
seed |
Random seed. |
within |
Optional grouping columns within which to permute. |
An R object containing permutation negative control. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Temporal-shift negative control
process_negative_control_shift(data, column, lag = 1L, by = NULL)process_negative_control_shift(data, column, lag = 1L, by = NULL)
data |
Data frame. |
column |
Column to shift. |
lag |
Number of rows to shift within each group. |
by |
Optional grouping columns. |
An R object containing temporal-shift negative control. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
N-gram features from process sequences
process_ngram_features(sequence, n = c(1L, 2L, 3L), separator = ">")process_ngram_features(sequence, n = c(1L, 2L, 3L), separator = ">")
sequence |
Sequence input. |
n |
Requested count or n-gram order, depending on context. |
separator |
Sequence-token separator. |
An object of class "matrix", stored as an R object, containing n-gram features from process sequences and associated metadata needed to interpret the result.
Compare an observed effect against a negative-control null distribution
process_null_benchmark(observed, controls, effect = "effect")process_null_benchmark(observed, controls, effect = "effect")
observed |
Observed scalar effect. |
controls |
Negative-control result or numeric vector. |
effect |
Effect column when controls is an object. |
A named list with components "observed", "n_null", "null_mean", "null_sd", "percentile", "two_sided_tail", "standardized_distance", containing an observed effect against a negative-control null distribution and associated metadata or diagnostics.
Produces model-discrepancy evidence; it never labels a participant as dishonest, impaired, disengaged, or otherwise psychologically categorized.
process_person_fit( object, data = NULL, person = NULL, response_weight = 1, rt_weight = 1, process_weight = 1 )process_person_fit( object, data = NULL, person = NULL, response_weight = 1, rt_weight = 1, process_weight = 1 )
object |
A fitted eyeprocess model or audit object. |
data |
Input data frame or compatible tabular object. |
person |
Person or participant identifier column. |
response_weight |
Weight assigned to response discrepancy. |
rt_weight |
Weight assigned to response-time discrepancy. |
process_weight |
Weight assigned to process discrepancy. |
An object of class "eye_process_person_fit", "data.frame", stored as a data frame, containing joint response-process person-fit diagnostic and associated metadata needed to interpret the result.
Defines transparent, review-oriented thresholds for incoming gaze/pupil data. The specification is a data-quality governance object, not a behavioral or clinical classifier.
process_preflight_spec( min_gaze_validity = 0.8, min_pupil_validity = 0.7, max_gaze_missingness = 0.25, max_pupil_missingness = 0.3, min_valid_trial_fraction = 0.7, trial_gaze_validity_threshold = 0.75, min_rt_ms = 200, max_rt_ms = 10000, sampling_rate_tolerance = 0.2, blink_quantile = 0.95, caution_flags = 1L, review_flags = 2L )process_preflight_spec( min_gaze_validity = 0.8, min_pupil_validity = 0.7, max_gaze_missingness = 0.25, max_pupil_missingness = 0.3, min_valid_trial_fraction = 0.7, trial_gaze_validity_threshold = 0.75, min_rt_ms = 200, max_rt_ms = 10000, sampling_rate_tolerance = 0.2, blink_quantile = 0.95, caution_flags = 1L, review_flags = 2L )
min_gaze_validity |
Minimum mean gaze-validity proportion. |
min_pupil_validity |
Minimum mean pupil-validity proportion. |
max_gaze_missingness |
Maximum mean gaze-missingness proportion. |
max_pupil_missingness |
Maximum mean pupil-missingness proportion. |
min_valid_trial_fraction |
Minimum fraction of trials meeting the trial-level gaze-validity threshold. |
trial_gaze_validity_threshold |
Gaze-validity threshold used to count an acceptable trial. |
min_rt_ms, max_rt_ms
|
Plausible mean response-time bounds in milliseconds. |
sampling_rate_tolerance |
Fractional deviation from the cohort median sampling rate that triggers review. |
blink_quantile |
Cohort quantile used for an extreme blink-cluster flag. |
caution_flags |
Number of flags yielding 'use_with_caution'. |
review_flags |
Number of flags yielding 'review_or_exclude_from_biometric_models'. |
An 'eye_process_preflight_spec' object.
Extract process-profile probabilities
process_profile_probabilities(x)process_profile_probabilities(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing process-profile probabilities. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarize process profiles
process_profile_summary(x)process_profile_summary(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing process profiles. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Test-retest process reliability profile
process_reliability_profile(data, person, session, measure)process_reliability_profile(data, person, session, measure)
data |
Long data. |
person, session, measure
|
Column names. |
An object of class "eye_process_reliability_profile", stored as a named list, with components "measure", "icc", "bland_altman", "caveat". It contains test-retest process reliability profile and associated metadata or diagnostics needed to interpret the result.
Return person/item latent-space coordinates
process_residual_map(object, entity = c("both", "person", "item"))process_residual_map(object, entity = c("both", "person", "item"))
object |
A fitted eyeprocess model or audit object. |
entity |
Entity type to map or validate. |
A tabular R object containing return person/item latent-space coordinates; rows represent analysis units and columns contain the returned quantities.
Construct an explicit process-analysis sensitivity grid
process_sensitivity_grid( ..., label = "process_sensitivity", max_specifications = 100000L )process_sensitivity_grid( ..., label = "process_sensitivity", max_specifications = 100000L )
... |
Named vectors of defensible analysis options. |
label |
Grid label. |
max_specifications |
Safety cap. |
A tabular R object containing an explicit process-analysis sensitivity grid; rows represent analysis units and columns contain the returned quantities.
Uses TF-IDF-weighted n-gram features and truncated SVD. This is a transparent classical embedding; sequence autoencoders can be supplied later through an external engine without changing the downstream contract.
process_sequence_embedding(sequence, n = c(1L, 2L, 3L), dimensions = 5L)process_sequence_embedding(sequence, n = c(1L, 2L, 3L), dimensions = 5L)
sequence |
Sequence input. |
n |
Requested count or n-gram order, depending on context. |
dimensions |
Number of embedding dimensions. |
An R object containing low-dimensional embedding of response-process sequences. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarize HMM state occupancy
process_state_occupancy(object)process_state_occupancy(object)
object |
A fitted eyeprocess model or audit object. |
An R object containing hMM state occupancy. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarize HMM process-state transitions
process_state_transition_summary(object)process_state_transition_summary(object)
object |
A fitted eyeprocess model or audit object. |
A data frame containing hMM process-state transitions. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Pairwise temporal stability across sessions
process_temporal_stability( data, person, session, measure, method = c("pearson", "spearman") )process_temporal_stability( data, person, session, measure, method = c("pearson", "spearman") )
data |
Long data. |
person, session, measure
|
Column names. |
method |
Correlation method. |
A logical value or vector indicating pairwise temporal stability across sessions.
Process measurement-uncertainty budgets. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
process_uncertainty_spec(calibration = TRUE, aoi_assignment = TRUE, preprocessing = TRUE, sampling = TRUE, model = TRUE, source_sd = NULL, draws = 1000, seed = 20260807) estimate_process_uncertainty(x, spec = process_uncertainty_spec(), metrics = NULL, cluster = NULL) propagate_process_uncertainty(x, estimand = function(data) mean(data, na.rm = TRUE), method = c("bootstrap", "simulation", "posterior"), draws = NULL, seed = NULL) uncertainty_budget(x) compare_uncertainty_budgets(...) plot_uncertainty_waterfall(x, ...) plot_uncertainty_tornado(x, ...) plot_uncertainty_by_item(x, ...) plot_uncertainty_by_stage(x, ...)process_uncertainty_spec(calibration = TRUE, aoi_assignment = TRUE, preprocessing = TRUE, sampling = TRUE, model = TRUE, source_sd = NULL, draws = 1000, seed = 20260807) estimate_process_uncertainty(x, spec = process_uncertainty_spec(), metrics = NULL, cluster = NULL) propagate_process_uncertainty(x, estimand = function(data) mean(data, na.rm = TRUE), method = c("bootstrap", "simulation", "posterior"), draws = NULL, seed = NULL) uncertainty_budget(x) compare_uncertainty_budgets(...) plot_uncertainty_waterfall(x, ...) plot_uncertainty_tornado(x, ...) plot_uncertainty_by_item(x, ...) plot_uncertainty_by_stage(x, ...)
calibration |
Argument controlling 'calibration'; see the function usage and returned audit metadata. |
aoi_assignment |
Argument controlling 'aoi_assignment'; see the function usage and returned audit metadata. |
preprocessing |
Argument controlling 'preprocessing'; see the function usage and returned audit metadata. |
sampling |
Argument controlling 'sampling'; see the function usage and returned audit metadata. |
model |
Argument controlling 'model'; see the function usage and returned audit metadata. |
source_sd |
Argument controlling 'source_sd'; see the function usage and returned audit metadata. |
draws |
Argument controlling 'draws'; see the function usage and returned audit metadata. |
seed |
Argument controlling 'seed'; see the function usage and returned audit metadata. |
x |
Input object or data structure appropriate for the selected analysis. |
spec |
Argument controlling 'spec'; see the function usage and returned audit metadata. |
metrics |
Argument controlling 'metrics'; see the function usage and returned audit metadata. |
cluster |
Argument controlling 'cluster'; see the function usage and returned audit metadata. |
estimand |
Argument controlling 'estimand'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
The design is intentionally explicit. It records measurement conditions under which recovery, uncertainty, convergence, and failure behavior will be evaluated. It does not imply that every combination is appropriate for every estimator.
process_validation_design( n_persons = c(50L, 150L, 500L), n_trials = c(10L, 30L, 80L), missingness = c(0, 0.05, 0.15, 0.3), sampling_rate_hz = c(60, 120, 300, 1000), aoi_error = c("low", "moderate", "severe"), calibration_error = c(0, 0.5, 1), pupil_dropout = c(0, 0.1, 0.3), heterogeneity = c("low", "moderate"), model_misspecification = c(FALSE, TRUE), replications = 100L, seed = 1L, label = "process_validation" )process_validation_design( n_persons = c(50L, 150L, 500L), n_trials = c(10L, 30L, 80L), missingness = c(0, 0.05, 0.15, 0.3), sampling_rate_hz = c(60, 120, 300, 1000), aoi_error = c("low", "moderate", "severe"), calibration_error = c(0, 0.5, 1), pupil_dropout = c(0, 0.1, 0.3), heterogeneity = c("low", "moderate"), model_misspecification = c(FALSE, TRUE), replications = 100L, seed = 1L, label = "process_validation" )
n_persons |
Participant counts. |
n_trials |
Trial/item counts per participant. |
missingness |
Proportion of generic process observations made missing. |
sampling_rate_hz |
Nominal sampling rates. |
aoi_error |
AOI uncertainty regimes. |
calibration_error |
Calibration-error regimes in user-defined units. |
pupil_dropout |
Pupil-specific dropout proportions. |
heterogeneity |
Participant-heterogeneity regimes. |
model_misspecification |
Logical regimes indicating deliberate mismatch. |
replications |
Monte Carlo replications per condition. |
seed |
Master simulation seed. |
label |
Optional design label. |
An 'eye_process_validation_design' object.
Specify temporal process windows
process_window_spec( width_ms = 1000, step_ms = 500, start_ms = 0, end_ms = 3000, align = c("stimulus", "response", "custom"), min_samples = 5L )process_window_spec( width_ms = 1000, step_ms = 500, start_ms = 0, end_ms = 3000, align = c("stimulus", "response", "custom"), min_samples = 5L )
width_ms |
Window width in milliseconds. |
step_ms |
Step between successive windows. |
start_ms, end_ms
|
Analysis range relative to the chosen alignment origin. |
align |
Alignment label, e.g. stimulus, response, or custom. |
min_samples |
Minimum samples required within a window. |
An 'eye_process_window_spec' object.
Promotion is an evidence record, not a mutable global status change unless 'update_registry = TRUE' is requested.
promote_irt_model( spec, evidence, target = c("experimental", "reference"), update_registry = FALSE )promote_irt_model( spec, evidence, target = c("experimental", "reference"), update_registry = FALSE )
spec |
IRT model or validation specification. |
evidence |
Validation evidence used for promotion. |
target |
Target evidence/status level. |
update_registry |
Whether the in-memory registry is updated. |
An object of class "eye_irt_promotion", stored as a named list, with components "model", "from", "to", "evidence_grade", "evidence_pass", "timestamp". It contains promote an IRT model after evidence gates are met and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
promote_vendor_support(corpus_path, case_id, level = c("fixture-tested", "empirically-validated"), validation, reviewer, notes = NA_character_)promote_vendor_support(corpus_path, case_id, level = c("fixture-tested", "empirically-validated"), validation, reviewer, notes = NA_character_)
corpus_path |
Corpus directory. |
case_id |
Case identifier. |
level |
New support level. |
validation |
Validation result/audit. |
reviewer |
Reviewer identifier. |
notes |
Notes. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Propagate empirical calibration uncertainty around gaze samples
propagate_calibration_uncertainty( data, model, x = "gaze_x", y = "gaze_y", draws = 500L, seed = 1L )propagate_calibration_uncertainty( data, model, x = "gaze_x", y = "gaze_y", draws = 500L, seed = 1L )
data |
Gaze samples. |
model |
Calibration error model. |
x, y
|
Gaze-coordinate columns. |
draws |
Monte Carlo draws per sample. |
seed |
Seed. |
An R object containing propagate empirical calibration uncertainty around gaze samples. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Build a provenance edge table
provenance_edge_table(from, to, relation = "wasDerivedFrom")provenance_edge_table(from, to, relation = "wasDerivedFrom")
from, to
|
Node identifiers. |
relation |
PROV-like relation labels. |
A data frame containing a provenance edge table. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Build a provenance lineage node table
provenance_lineage_table( id, type = "entity", label = id, value = NA_character_ )provenance_lineage_table( id, type = "entity", label = id, value = NA_character_ )
id |
Node identifiers. |
type |
Node types such as entity, activity, agent. |
label |
Human-readable labels. |
value |
Optional values/locations. |
A data frame containing a provenance lineage node table. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
prune_validation_checkpoints(path, statuses = c("corrupt", "locked"), dry_run = TRUE)prune_validation_checkpoints(path, statuses = c("corrupt", "locked"), dry_run = TRUE)
path |
Validation output directory. |
statuses |
Statuses to remove. |
dry_run |
Report without deleting. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Returns conservative metadata only. The package deliberately does not auto-download third-party human-participant data; users must review the source licence/terms and obtain data from the authoritative repository.
public_validation_corpus()public_validation_corpus()
A data frame containing public validation-corpus registry. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compute a transparent pupil activity index
pupil_activity_index( y, time_ms = seq_along(y), sampling_rate_hz = NULL, method = c("velocity", "frequency_contrast", "ripa_proxy"), low_band = c(0.05, 0.5), high_band = c(0.5, 4), fast_window_ms = 250, slow_window_ms = 750 )pupil_activity_index( y, time_ms = seq_along(y), sampling_rate_hz = NULL, method = c("velocity", "frequency_contrast", "ripa_proxy"), low_band = c(0.05, 0.5), high_band = c(0.5, 4), fast_window_ms = 250, slow_window_ms = 750 )
y |
Pupil signal. |
time_ms |
Time vector. |
sampling_rate_hz |
Sampling rate for frequency methods. |
method |
'velocity', 'frequency_contrast', or 'ripa_proxy'. |
low_band, high_band
|
Frequency bands for frequency contrast. |
fast_window_ms, slow_window_ms
|
Smoothing windows for the RIPA-style proxy. |
A numeric value or vector containing a transparent pupil activity index.
Compute pupil signal power in a frequency band
pupil_band_power(y, sampling_rate_hz, lower_hz, upper_hz, detrend = TRUE)pupil_band_power(y, sampling_rate_hz, lower_hz, upper_hz, detrend = TRUE)
y |
Pupil signal. |
sampling_rate_hz |
Sampling rate in Hz. |
lower_hz, upper_hz
|
Frequency-band limits. |
detrend |
Remove the mean before FFT. |
A numeric value or vector containing pupil signal power in a frequency band.
Evaluate pupil baseline-window sensitivity
pupil_baseline_sensitivity( data, time = "time_ms", pupil = "pupil", windows, by = NULL, correction = c("subtractive", "divisive") )pupil_baseline_sensitivity( data, time = "time_ms", pupil = "pupil", windows, by = NULL, correction = c("subtractive", "divisive") )
data |
Pupil data. |
time |
Time column. |
pupil |
Pupil column. |
windows |
Named list of two-element baseline windows. |
by |
Optional grouping columns. |
correction |
'subtractive' or 'divisive'. |
An R object containing pupil baseline-window sensitivity. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Extract pupil confound-model effects
pupil_confound_effects(x)pupil_confound_effects(x)
x |
Object to process, inspect, compare, or plot. |
A numeric coefficient table: the smooth-term table for 'mgcv::gam()' fits or the coefficient matrix for 'stats::lm()' fits.
Extract event effects from pupil deconvolution
pupil_event_effects(x)pupil_event_effects(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing event effects from pupil deconvolution. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Build an event-locked pupil regressor
pupil_event_regressor(time_ms, event_time_ms, tmax_ms = 930, shape = 10.1)pupil_event_regressor(time_ms, event_time_ms, tmax_ms = 930, shape = 10.1)
time_ms |
Sample times. |
event_time_ms |
Event onset. |
tmax_ms, shape
|
Kernel parameters. |
A numeric value or vector containing an event-locked pupil regressor.
Extract pupil frequency-domain and activity features by group
pupil_frequency_features( data, by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", sampling_rate_hz = 60, low_band = c(0.05, 0.5), high_band = c(0.5, 4) )pupil_frequency_features( data, by = c("person_id", "trial_id"), time = "time_ms", pupil = "pupil_bc", sampling_rate_hz = 60, low_band = c(0.05, 0.5), high_band = c(0.5, 4) )
data |
Sample-level data. |
by |
Grouping columns, e.g. person and trial/window. |
time, pupil
|
Column names. |
sampling_rate_hz |
Either a scalar or a column name. |
low_band, high_band
|
Frequency bands. |
An object of class "eye_pupil_frequency_features", stored as a named list, with components "features", "low_band", "high_band", "by", "pupil", "time", "caveat". It contains pupil frequency-domain and activity features by group and associated metadata or diagnostics needed to interpret the result.
Estimates pupil-response latency using a sustained threshold, maximum-slope tangent intersection, and piecewise breakpoint. The result reports estimator spread and a sampling/noise-aware resolvability label rather than treating a single latency estimate as algorithm- or hardware-independent.
pupil_latency_sensitivity( time, pupil, event_time = 0, baseline_window = c(-0.5, 0), search_window = c(0, 2), direction = c("constriction", "dilation"), threshold_sigma = 3, sustain_ms = 40 )pupil_latency_sensitivity( time, pupil, event_time = 0, baseline_window = c(-0.5, 0), search_window = c(0, 2), direction = c("constriction", "dilation"), threshold_sigma = 3, sustain_ms = 40 )
time |
Numeric sample times in seconds. |
pupil |
Numeric pupil values. |
event_time |
Nominal event time in seconds. |
baseline_window |
Two-element window relative to 'event_time'. |
search_window |
Two-element search window relative to 'event_time'. |
direction |
'"constriction"' or '"dilation"'. |
threshold_sigma |
Robust-noise multiples used by the sustained threshold. |
sustain_ms |
Required duration above threshold in milliseconds. |
A list of estimator-specific latencies, spread, signal diagnostics, resolvability, and provenance.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
pupil_preprocessing_grid(baseline_windows = list(c(-200, 0), c(-500, 0)), latency_ms = c(100, 200, 300), basis_df = c(4L, 6L, 8L), baseline_methods = c("subtract", "percent"), max_interpolated_fraction = c(0.10, 0.20))pupil_preprocessing_grid(baseline_windows = list(c(-200, 0), c(-500, 0)), latency_ms = c(100, 200, 300), basis_df = c(4L, 6L, 8L), baseline_methods = c("subtract", "percent"), max_interpolated_fraction = c(0.10, 0.20))
baseline_windows |
List of baseline windows. |
latency_ms |
Latency shifts. |
basis_df |
Basis degrees of freedom. |
baseline_methods |
Baseline corrections. |
max_interpolated_fraction |
Interpolation thresholds. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
pupil_preprocessing_sensitivity(x, grid = pupil_preprocessing_grid(), base_spec = functional_pupil_irt_spec(engine = "two_stage_glm"), fit = TRUE, extractor = extract_functional_pupil_parameters, continue_on_error = TRUE, ...)pupil_preprocessing_sensitivity(x, grid = pupil_preprocessing_grid(), base_spec = functional_pupil_irt_spec(engine = "two_stage_glm"), fit = TRUE, extractor = extract_functional_pupil_parameters, continue_on_error = TRUE, ...)
x |
Eye dataset or long pupil data. |
grid |
Sensitivity grid. |
base_spec |
Base functional pupil specification. |
fit |
Whether to fit each specification. |
extractor |
Optional result extractor. |
continue_on_error |
Record errors rather than stopping. |
... |
Passed to model fitting. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Canonical gamma-shaped pupil response kernel
pupil_response_kernel( time_since_event_ms, tmax_ms = 930, shape = 10.1, normalize = TRUE )pupil_response_kernel( time_since_event_ms, tmax_ms = 930, shape = 10.1, normalize = TRUE )
time_since_event_ms |
Time relative to event onset. |
tmax_ms |
Approximate response peak time. |
shape |
Shape parameter. |
normalize |
Normalize peak to one. |
A numeric value or vector containing canonical gamma-shaped pupil response kernel.
Pupil-unit semantic-fidelity audit
pupil_unit_fidelity_audit( source, roundtrip, source_pupil = "pupil_size", roundtrip_pupil = source_pupil, key = NULL, tolerance = 1e-06, correlation_floor = 0.995 )pupil_unit_fidelity_audit( source, roundtrip, source_pupil = "pupil_size", roundtrip_pupil = source_pupil, key = NULL, tolerance = 1e-06, correlation_floor = 0.995 )
source |
Original/source representation. |
roundtrip |
Round-tripped or comparison representation. |
source_pupil |
Source pupil-measure column. |
roundtrip_pupil |
Round-tripped pupil-measure column. |
key |
Column or columns used to align records. |
tolerance |
Numerical tolerance used by the comparison. |
correlation_floor |
Minimum correlation treated as compatible. |
An object of class "eye_pupil_fidelity", stored as a named list, with components "status", "matched_n", "correlation", "estimated_scale_ratio", "scaled_max_error", "tolerance". It contains pupil-unit semantic-fidelity audit and associated metadata or diagnostics needed to interpret the result.
Derivative-based pupil activity magnitude
pupil_velocity_activity(y, time_ms)pupil_velocity_activity(y, time_ms)
y |
Pupil signal. |
time_ms |
Time in milliseconds. |
A numeric value or vector containing derivative-based pupil activity magnitude.
Quantify leakage from row-wise rather than grouped validation
quantify_process_leakage( data, formula, group = c("participant_id", "item_id"), v = 5L, seed = 1L )quantify_process_leakage( data, formula, group = c("participant_id", "item_id"), v = 5L, seed = 1L )
data |
Data frame. |
formula |
Binary-outcome formula. |
group |
Grouping columns. |
v |
Number of folds. |
seed |
Random seed. |
Comparison table.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
query_eye_storage(storage, table, filters = list(), columns = NULL, collect = TRUE)query_eye_storage(storage, table, filters = list(), columns = NULL, collect = TRUE)
storage |
Storage object or path. |
table |
Canonical table. |
filters |
Named list of equality filters. |
columns |
Optional selected columns. |
collect |
Whether to collect an Arrow query. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Specify a published Raven strategy-model reproduction
raven_reproduction_spec( data_path, response, strategy_features, published_targets = NULL, licence_reviewed = FALSE, citation = "10.1016/j.intell.2023.101782" )raven_reproduction_spec( data_path, response, strategy_features, published_targets = NULL, licence_reviewed = FALSE, citation = "10.1016/j.intell.2023.101782" )
data_path |
Path to the exact public data/materials. |
response |
Response field. |
strategy_features |
Theory-defined eye-tracking strategy indicators. |
published_targets |
Optional named target estimates. |
licence_reviewed |
Whether data and code reuse has been reviewed. |
citation |
Citation or DOI for the reproduced analysis. |
An 'eye_raven_reproduction_spec'.
Read API lifecycle registry from CSV
read_api_lifecycle_registry(path)read_api_lifecycle_registry(path)
path |
CSV path. |
An R object containing aPI lifecycle registry from CSV. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
read_benchmark_table(study = eyeprocess_benchmark_study(), table)read_benchmark_table(study = eyeprocess_benchmark_study(), table)
study |
Benchmark object or path. |
table |
Table name. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Read a decision manifest written by eyeprocess
read_decision_manifest(path, format = NULL)read_decision_manifest(path, format = NULL)
path |
Input path. |
format |
Optional format; inferred from extension when omitted. |
A logical value or vector indicating a decision manifest written by eyeprocess.
Read and verify a frozen evidence bundle
read_eyeprocess_validation_evidence(path, verify = TRUE)read_eyeprocess_validation_evidence(path, verify = TRUE)
path |
File path for reading or writing. |
verify |
Whether integrity verification is performed when reading. |
An R object containing and verify a frozen evidence bundle. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Read a reproducibility fingerprint
read_reproducibility_fingerprint(path, format = NULL)read_reproducibility_fingerprint(path, format = NULL)
path |
Input path. |
format |
Optional format. |
An R object containing a reproducibility fingerprint. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
read_validation_job_manifest(path)read_validation_job_manifest(path)
path |
Manifest directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Read a validation scenario manifest
read_validation_scenario_manifest(path)read_validation_scenario_manifest(path)
path |
File path for reading or writing. |
An R object containing a validation scenario manifest. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
read_vendor_registry(corpus_path)read_vendor_registry(corpus_path)
corpus_path |
Corpus directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Iteratively detect, clean, and recalibrate after process change points
recalibrate_after_changepoint( data, fitter, person = "participant_id", order = "item_order", policy = c("flag", "exclude_post_change", "add_regime"), ... )recalibrate_after_changepoint( data, fitter, person = "participant_id", order = "item_order", policy = c("flag", "exclude_post_change", "add_regime"), ... )
data |
Input data frame or compatible tabular object. |
fitter |
Function accepting a data frame and returning a calibration fit. |
person |
Person or participant identifier column. |
order |
Within-sequence ordering variable. |
policy |
'flag', 'exclude_post_change', or 'add_regime'. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_changepoint_recalibration", stored as a named list, with components "changepoints", "data", "fit", "policy". It contains iteratively detect, clean, and recalibrate after process change points and associated metadata or diagnostics needed to interpret the result.
Approximate simulation replications needed for a target Monte Carlo error
recommended_validation_replications( target_mcse = 0.01, metric = c("coverage", "mean"), anticipated_sd = 1, anticipated_probability = 0.95, minimum = 100L )recommended_validation_replications( target_mcse = 0.01, metric = c("coverage", "mean"), anticipated_sd = 1, anticipated_probability = 0.95, minimum = 100L )
target_mcse |
Target Monte Carlo standard error. |
metric |
Metric to calculate or audit. |
anticipated_sd |
Anticipated standard deviation. |
anticipated_probability |
Anticipated probability for a binary metric. |
minimum |
Minimum acceptable value or threshold. |
A numeric value or vector containing approximate simulation replications needed for a target Monte Carlo error.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
redact_validation_case(source_path, output_path, id_columns = c("participant_id", "subject", "participant", "recording_id", "session_id"), remove_columns = c("name", "email", "address", "birthdate", "date_of_birth"), text_redactor = NULL, salt, copy_non_tabular = FALSE, overwrite = FALSE)redact_validation_case(source_path, output_path, id_columns = c("participant_id", "subject", "participant", "recording_id", "session_id"), remove_columns = c("name", "email", "address", "birthdate", "date_of_birth"), text_redactor = NULL, salt, copy_non_tabular = FALSE, overwrite = FALSE)
source_path |
Source file/directory. |
output_path |
Redacted output directory. |
id_columns |
Identifier columns to pseudonymize. |
remove_columns |
Columns to remove. |
text_redactor |
Optional function applied to character columns. |
salt |
Required project-specific salt. |
copy_non_tabular |
Copy unsupported files unchanged. |
overwrite |
Replace output. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Add or update API lifecycle metadata without global mutation
register_eye_api_status( registry = eye_api_lifecycle(), name, status, canonical = NA_character_, replacement = NA_character_, since = "0.9.0.9000", notes = NA_character_ )register_eye_api_status( registry = eye_api_lifecycle(), name, status, canonical = NA_character_, replacement = NA_character_, since = "0.9.0.9000", notes = NA_character_ )
registry |
Existing lifecycle registry. |
name |
API name. |
status |
Lifecycle status. |
canonical |
Canonical API for the same concept, if applicable. |
replacement |
Replacement for deprecated/superseded API. |
since |
Version in which status applies. |
notes |
Notes. |
An R object containing add or update API lifecycle metadata without global mutation. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Register a multimodal IRT model
register_irt_model(spec, overwrite = FALSE)register_irt_model(spec, overwrite = FALSE)
spec |
An 'irt_model_spec()'. |
overwrite |
Whether to replace an existing model with the same id. |
An R object containing a multimodal IRT model. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Add a process measure to a registry without global mutation
register_process_measure( registry = process_measure_registry(), name, channel, unit, level, interpretation, guardrail, status = "user_defined" )register_process_measure( registry = process_measure_registry(), name, channel, unit, level, interpretation, guardrail, status = "user_defined" )
registry |
Registry. |
name, channel, unit, level, interpretation, guardrail, status
|
Measure metadata. |
A tabular R object containing add a process measure to a registry without global mutation; rows represent analysis units and columns contain the returned quantities.
Pupil phase-amplitude registration. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
register_pupil_curves(x, time, pupil, anchor = c("stimulus", "response", "event"), method = c("elastic", "landmark"), id_col = "person_id", grid_size = 101) decompose_pupil_phase_amplitude(x, components = 3) fit_phase_amplitude_irt(responses, phase_scores, amplitude_scores = NULL, person_id = NULL, family = c("gaussian", "binomial"), ...) audit_pupil_registration(x) plot_pupil_registration(x, ...) plot_warping_functions(x, ...) plot_phase_amplitude_scores(x, ...) plot_item_phase_delay(x, ...) plot_registered_pupil_effects(x, ...)register_pupil_curves(x, time, pupil, anchor = c("stimulus", "response", "event"), method = c("elastic", "landmark"), id_col = "person_id", grid_size = 101) decompose_pupil_phase_amplitude(x, components = 3) fit_phase_amplitude_irt(responses, phase_scores, amplitude_scores = NULL, person_id = NULL, family = c("gaussian", "binomial"), ...) audit_pupil_registration(x) plot_pupil_registration(x, ...) plot_warping_functions(x, ...) plot_phase_amplitude_scores(x, ...) plot_item_phase_delay(x, ...) plot_registered_pupil_effects(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
time |
Argument controlling 'time'; see the function usage and returned audit metadata. |
pupil |
Argument controlling 'pupil'; see the function usage and returned audit metadata. |
anchor |
Argument controlling 'anchor'; see the function usage and returned audit metadata. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
id_col |
Argument controlling 'id_col'; see the function usage and returned audit metadata. |
grid_size |
Argument controlling 'grid_size'; see the function usage and returned audit metadata. |
components |
Argument controlling 'components'; see the function usage and returned audit metadata. |
responses |
Argument controlling 'responses'; see the function usage and returned audit metadata. |
phase_scores |
Argument controlling 'phase_scores'; see the function usage and returned audit metadata. |
amplitude_scores |
Argument controlling 'amplitude_scores'; see the function usage and returned audit metadata. |
person_id |
Argument controlling 'person_id'; see the function usage and returned audit metadata. |
family |
Argument controlling 'family'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
register_validation_case(corpus_path, source_path, vendor, device_model, software_name, software_version, hardware_version = NA_character_, export_profile = NA_character_, sampling_rate_hz = NA_real_, coordinate_system = NA_character_, timebase = NA_character_, event_semantics = NA_character_, ocular_structure = NA_character_, missingness_convention = NA_character_, vendor_fixations = NA_character_, package_transformations = NA_character_, unsupported_fields = NA_character_, independent_source = TRUE, licence_reviewed = FALSE, redistribution_allowed = FALSE, support_level = c("declared", "fixture-tested", "empirically-validated"), mode = c("reference", "copy"), case_id = NULL, notes = NA_character_)register_validation_case(corpus_path, source_path, vendor, device_model, software_name, software_version, hardware_version = NA_character_, export_profile = NA_character_, sampling_rate_hz = NA_real_, coordinate_system = NA_character_, timebase = NA_character_, event_semantics = NA_character_, ocular_structure = NA_character_, missingness_convention = NA_character_, vendor_fixations = NA_character_, package_transformations = NA_character_, unsupported_fields = NA_character_, independent_source = TRUE, licence_reviewed = FALSE, redistribution_allowed = FALSE, support_level = c("declared", "fixture-tested", "empirically-validated"), mode = c("reference", "copy"), case_id = NULL, notes = NA_character_)
corpus_path |
Corpus directory. |
source_path |
Real export file or directory. |
vendor |
Vendor name. |
device_model |
Hardware model. |
software_name |
Export software and version. |
software_version |
Export software and version. |
hardware_version |
Optional hardware/firmware version. |
export_profile |
Export options/profile. |
sampling_rate_hz |
Nominal or observed rate. |
coordinate_system |
Semantics metadata. |
timebase |
Semantics metadata. |
event_semantics |
Semantics metadata. |
ocular_structure |
Monocular/binocular structure. |
missingness_convention |
Vendor missing-value convention. |
vendor_fixations |
Description of vendor-derived fixation fields. |
package_transformations |
Declared package transformations. |
unsupported_fields |
Known unsupported fields. |
independent_source |
Whether independently obtained. |
licence_reviewed |
Whether licensing review is complete. |
redistribution_allowed |
Whether redacted material may be redistributed. |
support_level |
Declared support level. |
mode |
Reference source in place or copy it into the private corpus. |
case_id |
Optional explicit identifier. |
notes |
Notes. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
register_vendor_semantics(corpus_path, vendor, native_field, native_meaning, canonical_table, canonical_field, unit = NA_character_, transformation = "identity", loss_risk = c("none", "low", "moderate", "high", "unsupported"), evidence_case_id = NA_character_)register_vendor_semantics(corpus_path, vendor, native_field, native_meaning, canonical_table, canonical_field, unit = NA_character_, transformation = "identity", loss_risk = c("none", "low", "moderate", "high", "unsupported"), evidence_case_id = NA_character_)
corpus_path |
Corpus directory. |
vendor |
Vendor. |
native_field |
Native field and meaning. |
native_meaning |
Native field and meaning. |
canonical_table |
Canonical destination. |
canonical_field |
Canonical destination. |
unit |
Unit. |
transformation |
Transformation description. |
loss_risk |
None, low, moderate, high, or unsupported. |
evidence_case_id |
Supporting case. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Audit reporting-guideline coverage
reporting_guideline_audit(x, model = NULL, sensitivity = NULL)reporting_guideline_audit(x, model = NULL, sensitivity = NULL)
x |
An 'eye_dataset'. |
model |
Optional 'eyeprocess_model'. |
sensitivity |
Optional 'eye_multiverse' sensitivity result. |
An 'eye_reporting_audit' data frame.
Representative scanpaths and scanpath distributions. These functions form the eyeprocess 0.6.0.9000 measurement-intelligence programme and use dependency-free reference implementations with explicit evidence limits.
representative_scanpath(x, method = c("medoid", "barycenter", "consensus"), id_col = "person_id", aoi_col = "aoi", x_col = "x", y_col = "y", distance = c("multimatch", "edit", "transport")) scanpath_dispersion(x) compare_scanpath_distributions(x, group, distance = c("multimatch", "edit", "transport"), permutations = 499) bootstrap_representative_scanpath(x, draws = 250, seed = 20260807) plot_scanpath_atlas(x, ...) plot_representative_scanpath(x, ...) plot_scanpath_dispersion(x, ...) plot_group_scanpath_transport(x, ...) plot_scanpath_similarity_matrix(x, ...)representative_scanpath(x, method = c("medoid", "barycenter", "consensus"), id_col = "person_id", aoi_col = "aoi", x_col = "x", y_col = "y", distance = c("multimatch", "edit", "transport")) scanpath_dispersion(x) compare_scanpath_distributions(x, group, distance = c("multimatch", "edit", "transport"), permutations = 499) bootstrap_representative_scanpath(x, draws = 250, seed = 20260807) plot_scanpath_atlas(x, ...) plot_representative_scanpath(x, ...) plot_scanpath_dispersion(x, ...) plot_group_scanpath_transport(x, ...) plot_scanpath_similarity_matrix(x, ...)
x |
Input object or data structure appropriate for the selected analysis. |
method |
Argument controlling 'method'; see the function usage and returned audit metadata. |
id_col |
Argument controlling 'id_col'; see the function usage and returned audit metadata. |
aoi_col |
Argument controlling 'aoi_col'; see the function usage and returned audit metadata. |
x_col |
Argument controlling 'x_col'; see the function usage and returned audit metadata. |
y_col |
Argument controlling 'y_col'; see the function usage and returned audit metadata. |
distance |
Argument controlling 'distance'; see the function usage and returned audit metadata. |
group |
Argument controlling 'group'; see the function usage and returned audit metadata. |
permutations |
Argument controlling 'permutations'; see the function usage and returned audit metadata. |
draws |
Argument controlling 'draws'; see the function usage and returned audit metadata. |
seed |
Argument controlling 'seed'; see the function usage and returned audit metadata. |
... |
Additional arguments passed to the underlying method or plotting function. |
The APIs return auditable S3 objects. Plot wrappers call registered base-graphics methods. Experimental or approximate engines are labelled in object status fields and should be validated before confirmatory or operational use.
An eyeprocess result object, data frame, model object, plot, or audit table as documented by the individual function.
plot_diagnostics(), plot_evidence(), and plot_sensitivity().
Resume a governed pipeline from a prior run
resume_eye_pipeline(x, previous, context = list(), stop_on_error = TRUE)resume_eye_pipeline(x, previous, context = list(), stop_on_error = TRUE)
x |
Pipeline. |
previous |
Prior pipeline run. |
context |
Context used for new steps. |
stop_on_error |
Stop on non-optional error. |
An object of class "eye_pipeline_run", stored as a named list, with components "pipeline", "pipeline_hash", "outputs", "records", "errors", "warnings", "context_hash", "completed", "created_at", "status". It contains resume a governed pipeline from a prior run and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
resume_validation_jobs(plan, output_dir, retry = c("missing", "failed", "nonconverged", "locked", "corrupt"), ...)resume_validation_jobs(plan, output_dir, retry = c("missing", "failed", "nonconverged", "locked", "corrupt"), ...)
plan |
Validation plan or manifest path. |
output_dir |
Validation output directory. |
retry |
Which statuses to re-run. |
... |
Passed to 'run_validation_jobs()'. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
This is a callback harness so it remains stable even if the package's BIDS writer/reader signatures evolve. 'exporter' receives the source object plus 'export_args'; 'importer' receives the exporter result plus 'import_args'.
roundtrip_eye_bids( source, exporter, importer, export_args = list(), import_args = list(), extract_samples = .ep07_roundtrip_extract_samples, audit_args = list() )roundtrip_eye_bids( source, exporter, importer, export_args = list(), import_args = list(), extract_samples = .ep07_roundtrip_extract_samples, audit_args = list() )
source |
Source eyeprocess object/table. |
exporter |
Function that writes/exports BIDS and returns a locator or object consumable by 'importer'. |
importer |
Function that reconstructs an eyeprocess object/table. |
export_args, import_args
|
Named argument lists. |
extract_samples |
Function extracting the canonical sample table from source and reconstructed objects. |
audit_args |
Arguments forwarded to 'semantic_roundtrip_audit()'. |
An object of class "eye_bids_roundtrip", stored as a named list, with components "exported", "reconstructed", "audit", "status". It contains execute and audit an Eye-Tracking-BIDS round trip and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
run_benchmark_reproduction(study = eyeprocess_benchmark_study())run_benchmark_reproduction(study = eyeprocess_benchmark_study())
study |
Benchmark object or path. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Run a computational scaling benchmark
run_eye_benchmark( design = eye_benchmark_design(), generator = .ep09_default_benchmark_generator, operation = .ep09_default_benchmark_operation, gc_before = TRUE, progress = interactive() )run_eye_benchmark( design = eye_benchmark_design(), generator = .ep09_default_benchmark_generator, operation = .ep09_default_benchmark_operation, gc_before = TRUE, progress = interactive() )
design |
Benchmark design. |
generator |
Function '(n, row)' returning benchmark input. |
operation |
Function '(data, row)' representing the operation under test. |
gc_before |
Run garbage collection before timing. |
progress |
Print progress. |
An object of class "eye_benchmark_result", stored as a named list, with components "design", "results", "created_at", "status", "caveat". It contains a computational scaling benchmark and associated metadata or diagnostics needed to interpret the result.
Run a governed eyeprocess pipeline
run_eye_pipeline(x, context = list(), stop_on_error = TRUE, previous = NULL)run_eye_pipeline(x, context = list(), stop_on_error = TRUE, previous = NULL)
x |
Pipeline. |
context |
Initial named context available as '.context'. |
stop_on_error |
Stop on a non-optional step error. |
previous |
Optional prior 'eye_pipeline_run' used for resumption. |
An object of class "eye_pipeline_run", stored as a named list, with components "pipeline", "pipeline_hash", "outputs", "records", "errors", "warnings", "context_hash", "completed", "created_at", "status". It contains a governed eyeprocess pipeline and associated metadata or diagnostics needed to interpret the result.
The caller supplies an exported equateIRT function name and its arguments. eyeprocess does not replace the requested equating estimator when the engine is unavailable.
run_eyeprocess_equateirt(function_name, ..., engine = "equateIRT")run_eyeprocess_equateirt(function_name, ..., engine = "equateIRT")
function_name |
Name of the external equateIRT function to call. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "function_name", "fit", "call". It contains a named equateIRT linking/equating function without fallback substitution and associated metadata or diagnostics needed to interpret the result.
Simulates abilities from the declared normal prior, responses from the known item-response model, and posterior draws from the same grid-based scoring algorithm used by 'eyeprocess_irt_eap_score()'. This validates computational calibration of the scoring workflow under the declared generative model; it does not establish empirical adequacy or construct validity.
run_eyeprocess_irt_ability_sbc( items, replications = 200L, posterior_draws = 99L, theta_grid = seq(-5, 5, length.out = 401), prior_mean = 0, prior_sd = 1, interval = 0.95, seed = 20260811L, D = 1 )run_eyeprocess_irt_ability_sbc( items, replications = 200L, posterior_draws = 99L, theta_grid = seq(-5, 5, length.out = 401), prior_mean = 0, prior_sd = 1, interval = 0.95, seed = 20260811L, D = 1 )
items |
Item-parameter data frame or item collection. |
replications |
Number of simulation or validation replications. |
posterior_draws |
Number of posterior draws generated per SBC replication. |
theta_grid |
Grid of latent-trait values used for numerical scoring or integration. |
prior_mean |
Mean of the normal latent-trait prior. |
prior_sd |
Standard deviation of the normal latent-trait prior. |
interval |
Central posterior interval probability used for coverage assessment. |
seed |
Random-number seed for reproducible execution. |
D |
Logistic scaling constant. |
An object of class "eye_irt_sbc_evidence", stored as a named list, with components "diagnostics", "ecdf_deviation", "n", "n_draws". It contains simulation-based calibration for known-item IRT ability scoring and associated metadata or diagnostics needed to interpret the result.
When mirt is unavailable the function returns a gated result rather than a substitute estimator.
run_eyeprocess_irt_recovery(design, engine = "mirt", verbose = TRUE)run_eyeprocess_irt_recovery(design, engine = "mirt", verbose = TRUE)
design |
Validation or simulation design object. |
engine |
Requested estimation or analysis engine. |
verbose |
Value supplied for the verbose argument. |
An object of class "eye_irt_recovery_result", stored as a named list, with components "design", "estimates", "failures", "engine". It contains iRT parameter recovery with the exact mirt engine and associated metadata or diagnostics needed to interpret the result.
Run a mirtCAT adaptive-testing workflow without fallback substitution
run_eyeprocess_mirtcat(..., engine = "mirtCAT")run_eyeprocess_mirtcat(..., engine = "mirtCAT")
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains a mirtCAT adaptive-testing workflow without fallback substitution and associated metadata or diagnostics needed to interpret the result.
'corruptors' is a named list of functions accepting '(data, severity, seed)'. 'metric_fun' must return a named finite/numeric vector (NA is allowed for metrics that are undefined in a scenario). The executor records software behavior under declared corruptions and does not define universal data- quality thresholds.
run_eyeprocess_stress_evidence(data, plan, corruptors, metric_fun)run_eyeprocess_stress_evidence(data, plan, corruptors, metric_fun)
data |
Input data frame, matrix, or compatible analysis object. |
plan |
Validation or stress-evidence plan object. |
corruptors |
Named list of corruption functions used by the stress programme. |
metric_fun |
Function used to compute the stress-programme evaluation metric. |
An object of class "eye_stress_evidence_result", stored as a named list, with components "plan", "scenarios", "baseline", "results", "failures", "guardrail". It contains execute a declared measurement-stress evidence plan and associated metadata or diagnostics needed to interpret the result.
Run the complete validation-release programme
run_eyeprocess_validation_program( corpus, output_dir, model_jobs = list(), sbc_jobs = list(), engine_jobs = list(), reproduction_jobs = list(), grouped_jobs = list(), leakage_jobs = list(), multiverse_jobs = list(), benchmark_jobs = list(), reporting_dataset = NULL, public_benchmark_dataset = NULL, public_benchmark_include_samples = FALSE, advanced_evidence = list(), evidence_spec = advanced_model_evidence_spec(), overwrite = FALSE )run_eyeprocess_validation_program( corpus, output_dir, model_jobs = list(), sbc_jobs = list(), engine_jobs = list(), reproduction_jobs = list(), grouped_jobs = list(), leakage_jobs = list(), multiverse_jobs = list(), benchmark_jobs = list(), reporting_dataset = NULL, public_benchmark_dataset = NULL, public_benchmark_include_samples = FALSE, advanced_evidence = list(), evidence_spec = advanced_model_evidence_spec(), overwrite = FALSE )
corpus |
Validation corpus or manifest. |
output_dir |
Output directory. |
model_jobs |
Named list of model-validation job specifications. Each job supplies 'simulator', 'fitter', 'extractor', 'truth_extractor', and optional 'grid' and 'spec'. |
sbc_jobs |
Named list of simulation-based-calibration job specifications. |
engine_jobs |
Named list of equivalent-engine comparison jobs. |
reproduction_jobs |
Named list of licensed empirical-reproduction jobs. |
grouped_jobs |
Named list of grouped-validation jobs. Set 'crossed = TRUE' in a job to call 'crossed_grouped_cv()'. |
leakage_jobs |
Named list of leakage-quantification jobs. |
multiverse_jobs |
Named list of preprocessing-multiverse jobs. |
benchmark_jobs |
Named list of zero-argument functions or benchmark argument lists. |
reporting_dataset |
Optional 'eye_dataset' for reporting-guideline coverage. |
public_benchmark_dataset |
Optional 'eye_dataset' from which to write a de-identified public benchmark bundle. |
public_benchmark_include_samples |
Whether the public benchmark retains sample-level tables. |
advanced_evidence |
Optional named evidence records keyed by model function. |
evidence_spec |
Advanced-model promotion-gate specification. |
overwrite |
Whether to replace the output directory. |
An 'eye_validation_program' object.
Imports a real Gazepoint folder, constructs a canonical 'eye_dataset', runs QC, reconstructs media trials, processes pupil and biometric streams, derives gaze/AOI/pupil/biometric features, creates analysis and IRT tables, generates plots, exports every stage, and writes a reproducible report.
run_gazepoint_workflow( path, output_dir = file.path(getwd(), "eyeprocess-gazepoint-workflow"), responses = NULL, score_key = NULL, item_map = NULL, spec = gazepoint_workflow_spec(), overwrite = FALSE, quiet = FALSE )run_gazepoint_workflow( path, output_dir = file.path(getwd(), "eyeprocess-gazepoint-workflow"), responses = NULL, score_key = NULL, item_map = NULL, spec = gazepoint_workflow_spec(), overwrite = FALSE, quiet = FALSE )
path |
Gazepoint export folder. |
output_dir |
Destination directory. |
responses |
Optional response data frame or CSV path. |
score_key |
Optional named vector of correct responses by item id. |
item_map |
Optional stimulus-to-item map. |
spec |
Workflow specification from 'gazepoint_workflow_spec()'. |
overwrite |
Replace an existing output directory. |
quiet |
Suppress progress messages. |
An 'eye_gazepoint_workflow' object.
The fitter may return a model or throw an error. The extractor must return a data frame with 'parameter', 'estimate', and optionally 'std_error', 'lower', and 'upper'. The truth extractor must return a named numeric vector.
run_model_validation( simulator, fitter, extractor, truth_extractor, grid = NULL, spec = model_validation_spec(), seed = 1L, continue_on_error = TRUE )run_model_validation( simulator, fitter, extractor, truth_extractor, grid = NULL, spec = model_validation_spec(), seed = 1L, continue_on_error = TRUE )
simulator |
Simulation function. |
fitter |
Estimation function receiving the simulation result. |
extractor |
Parameter extraction function. |
truth_extractor |
Truth extraction function. |
grid |
Scenario grid. |
spec |
Validation specification. |
seed |
Random seed. |
continue_on_error |
Record rather than stop on estimation errors. |
An 'eye_model_validation' object.
Run posterior simulation-based calibration from an explicit contract
run_posterior_sbc( observed_data, contract, replications = 100L, seed = 20260808L )run_posterior_sbc( observed_data, contract, replications = 100L, seed = 20260808L )
observed_data |
Observed dataset used by the calibration procedure. |
contract |
Posterior-SBC or validation contract. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_posterior_sbc", "eye_irt_sbc", stored as a named list, with components "ranks", "failures", "replications", "seed", "method", "requirement". It contains posterior simulation-based calibration from an explicit contract and associated metadata or diagnostics needed to interpret the result.
Run repeated process negative controls
run_process_negative_controls( data, outcome, analysis_fun, controls = c("permutation", "shift"), replications = 100L, seed = 1L, extract_fun = .ep09_default_control_extract, shift_lags = c(-3L, -2L, -1L, 1L, 2L, 3L), within = NULL )run_process_negative_controls( data, outcome, analysis_fun, controls = c("permutation", "shift"), replications = 100L, seed = 1L, extract_fun = .ep09_default_control_extract, shift_lags = c(-3L, -2L, -1L, 1L, 2L, 3L), within = NULL )
data |
Data frame. |
outcome |
Outcome column. |
analysis_fun |
Function applied to each negative-control dataset. |
controls |
Character vector among 'permutation' and 'shift'. |
replications |
Number of controls per type. |
seed |
Seed. |
extract_fun |
Function converting analysis result to a data.frame. |
shift_lags |
Lags sampled for shift controls. |
within |
Optional permutation groups. |
A named list with components "results", "outcome", "controls", "replications", "seed", "interpretation", containing repeated process negative controls and associated metadata or diagnostics.
Run an explicit process-analysis multiverse
run_process_sensitivity( data, grid, analysis_fun, extract_fun = .ep09_default_sensitivity_extract, progress = interactive() )run_process_sensitivity( data, grid, analysis_fun, extract_fun = .ep09_default_sensitivity_extract, progress = interactive() )
data |
Analysis data. |
grid |
Sensitivity grid. |
analysis_fun |
Function '(data, specification)'. |
extract_fun |
Function '(fit, specification)' returning one or more rows. |
progress |
Print progress. |
An object of class "eye_process_sensitivity", stored as a named list, with components "grid", "results", "failures", "warnings", "grid_hash", "created_at", "status", "caveat". It contains an explicit process-analysis multiverse and associated metadata or diagnostics needed to interpret the result.
Run an empirical process-validation programme
run_process_validation( design, simulate_fun = simulate_process_validation_data, fit_fun = .ep09_default_validation_fit, extract_fun = .ep09_default_validation_extract, max_conditions = Inf, progress = interactive() )run_process_validation( design, simulate_fun = simulate_process_validation_data, fit_fun = .ep09_default_validation_fit, extract_fun = .ep09_default_validation_extract, max_conditions = Inf, progress = interactive() )
design |
Validation design or expanded condition table. |
simulate_fun |
Function '(condition, replication, seed)' returning a simulation object. |
fit_fun |
Function '(simulated, condition)' returning a fitted object. |
extract_fun |
Function '(fit, simulated, condition)' returning one or more rows with estimates. |
max_conditions |
Optional cap on conditions actually run. |
progress |
Print compact progress messages. |
An 'eye_process_validation_result' object.
Execute a licensed published-model reproduction
run_raven_reproduction(spec, importer, fitter, extractor, tolerance = 0.05)run_raven_reproduction(spec, importer, fitter, extractor, tolerance = 0.05)
spec |
Reproduction specification. |
importer |
Function receiving 'spec$data_path'. |
fitter |
Function receiving imported data and 'spec'. |
extractor |
Function returning named estimates or a parameter/estimate data frame. |
tolerance |
Absolute target tolerance. |
An 'eye_empirical_reproduction' object.
Run generic simulation-based calibration
run_sbc( simulator, fitter, posterior_draws, replications = 100L, seed = 20260808L )run_sbc( simulator, fitter, posterior_draws, replications = 100L, seed = 20260808L )
simulator |
Function 'simulator(replicate)' returning 'list(data, truth)'; 'truth' must be a named numeric vector. |
fitter |
Function 'fitter(data)' returning a fitted object. |
posterior_draws |
Function returning a draws matrix/data frame whose columns match names in 'truth'. |
replications |
Number of SBC replications. |
seed |
RNG seed. |
An object of class "eye_irt_sbc", stored as a named list, with components "ranks", "failures", "replications", "seed", "method". It contains generic simulation-based calibration and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
run_validation_jobs(plan, simulator, fitter, extractor, truth_extractor, output_dir, workers = 1L, backend = c("auto", "sequential", "future"), isolation = c("auto", "in_process", "callr"), timeout_seconds = Inf, memory_limit_mb = Inf, stale_lock_seconds = 3600, overwrite = FALSE, fail_fast = FALSE, progress = interactive(), job_ids = NULL, chunks = NULL, simulation_args = list(), fit_args = list(), diagnostics_extractor = NULL, draws_extractor = NULL, predictions_extractor = NULL, confidence = 0.95, run_metadata = list())run_validation_jobs(plan, simulator, fitter, extractor, truth_extractor, output_dir, workers = 1L, backend = c("auto", "sequential", "future"), isolation = c("auto", "in_process", "callr"), timeout_seconds = Inf, memory_limit_mb = Inf, stale_lock_seconds = 3600, overwrite = FALSE, fail_fast = FALSE, progress = interactive(), job_ids = NULL, chunks = NULL, simulation_args = list(), fit_args = list(), diagnostics_extractor = NULL, draws_extractor = NULL, predictions_extractor = NULL, confidence = 0.95, run_metadata = list())
plan |
Validation plan or manifest directory. |
simulator |
Simulation function. |
fitter |
Fitting function receiving the simulated object first. |
extractor |
Function extracting parameter estimates. |
truth_extractor |
Function extracting named true parameter values. |
output_dir |
Validation output directory. |
workers |
Number of workers. |
backend |
Sequential or optional 'future' backend. |
isolation |
In-process execution or optional 'callr' isolation. |
timeout_seconds |
Per-job timeout. Enforced only with 'callr' isolation. |
memory_limit_mb |
Best-effort per-job memory limit. |
stale_lock_seconds |
Age after which an abandoned job lock may be reclaimed. |
overwrite |
Re-run completed checkpoints. |
fail_fast |
Stop after the first failed/nonconverged job. |
progress |
Display progress in sequential mode. |
job_ids |
Optional subset of job identifiers. |
chunks |
Optional subset of chunk identifiers. |
simulation_args |
Additional simulator arguments. |
fit_args |
Additional fitter arguments. |
diagnostics_extractor |
Optional diagnostics extractor. |
draws_extractor |
Optional posterior-draw extractor. |
predictions_extractor |
Optional prediction extractor. |
confidence |
Confidence level used when standard errors are supplied. |
run_metadata |
Named metadata included in the runner fingerprint; use it to record code, prior, or engine variants captured outside function bodies. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
ECDF deviation summary for SBC ranks
sbc_ecdf_deviation(x, n_draws = NULL)sbc_ecdf_deviation(x, n_draws = NULL)
x |
SBC diagnostics or ranks. |
n_draws |
Required if x is ranks. |
An R object containing eCDF deviation summary for SBC ranks. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Build SBC rank diagnostics
sbc_rank_diagnostics(ranks, n_draws, bins = NULL)sbc_rank_diagnostics(ranks, n_draws, bins = NULL)
ranks |
Integer rank statistics from 0 through 'n_draws'. |
n_draws |
Number of posterior draws used per rank. |
bins |
Histogram bins; defaults to a bounded square-root rule. |
'eye_sbc_diagnostics' object.
Summarize simulation-based calibration
sbc_summary(x)sbc_summary(x)
x |
An 'eye_sbc' object. |
Parameter-level rank and standardized-bias summaries.
Score a partial response pattern from a calibrated mirt model
score_partial_response_pattern( model, response_pattern, method = c("MAP", "EAP"), ... )score_partial_response_pattern( model, response_pattern, method = c("MAP", "EAP"), ... )
model |
Calibrated 'mirt' model. |
response_pattern |
Full-length response vector with future/unobserved items as NA. |
method |
mirt scoring method, typically MAP or EAP. |
... |
Passed to 'mirt::fscores()'. |
A data frame containing a partial response pattern from a calibrated mirt model. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Score a response stream cumulatively
score_response_stream( model, response_pattern, observed_order = NULL, method = c("MAP", "EAP"), ... )score_response_stream( model, response_pattern, observed_order = NULL, method = c("MAP", "EAP"), ... )
model |
Calibrated 'mirt' model. |
response_pattern |
Complete or partial response vector in item order. |
observed_order |
Order in which observed items arrive. Defaults to sequence. |
method |
MAP or EAP. |
... |
Passed to 'mirt::fscores()'. |
An 'eye_streaming_score' object.
Select the next item using response/process utility
select_next_item_process( theta, item_bank, used = character(), weights = c(response = 1, rt = 0, process = 0), burden_weight = 0 )select_next_item_process( theta, item_bank, used = character(), weights = c(response = 1, rt = 0, process = 0), burden_weight = 0 )
theta |
Latent-trait values. |
item_bank |
Value supplied to 'item_bank'; see Details for its model-specific role. |
used |
Items already used or unavailable for selection. |
weights |
Weights used to combine information components. |
burden_weight |
Penalty applied to expected burden. |
A named list with components "item_id", "utility", "row", "all_utilities", containing the next item using response/process utility and associated metadata or diagnostics.
Semantic fidelity specification
semantic_fidelity_spec( timestamp_tolerance = 1e-06, coordinate_tolerance = 1e-06, pupil_tolerance = 1e-06, missingness_tolerance = 1e-06, correlation_floor = 0.999, allow_row_reorder = TRUE )semantic_fidelity_spec( timestamp_tolerance = 1e-06, coordinate_tolerance = 1e-06, pupil_tolerance = 1e-06, missingness_tolerance = 1e-06, correlation_floor = 0.999, allow_row_reorder = TRUE )
timestamp_tolerance |
Absolute tolerance after time-unit normalization. |
coordinate_tolerance |
Absolute tolerance after coordinate normalization. |
pupil_tolerance |
Absolute tolerance after pupil-unit normalization. |
missingness_tolerance |
Maximum tolerated absolute change in missingness. |
correlation_floor |
Correlation floor used when deciding whether a numeric transformation remains semantically equivalent. |
allow_row_reorder |
Whether row reordering is permitted when a key is supplied. |
An object of class 'eye_semantic_fidelity_spec'.
Convert a semantic round-trip audit into a loss map
semantic_loss_map(x)semantic_loss_map(x)
x |
Object to print, plot, summarize, or audit. |
A tabular R object containing a semantic round-trip audit into a loss map; rows represent analysis units and columns contain the returned quantities.
Audit a complete semantic round trip
semantic_roundtrip_audit( source, roundtrip, key = NULL, fields = NULL, timestamp = list(), coordinates = list(), pupil = NULL, eye = NULL, source_events = NULL, roundtrip_events = NULL, event_args = list() )semantic_roundtrip_audit( source, roundtrip, key = NULL, fields = NULL, timestamp = list(), coordinates = list(), pupil = NULL, eye = NULL, source_events = NULL, roundtrip_events = NULL, event_args = list() )
source |
Original canonical samples. |
roundtrip |
Canonical samples reconstructed after interchange. |
key |
Optional row identity key. |
fields |
Fields for field-level comparison. |
timestamp |
Optional list of arguments forwarded to 'timestamp_fidelity_audit()'. |
coordinates |
Optional list of arguments forwarded to 'coordinate_fidelity_audit()'. |
pupil |
Optional list of arguments forwarded to 'pupil_unit_fidelity_audit()'. |
eye |
Optional list of arguments forwarded to 'eye_stream_fidelity_audit()'. |
source_events, roundtrip_events
|
Optional event tables. |
event_args |
Optional event-audit arguments. |
An 'eye_semantic_roundtrip' object.
Stable fingerprint of a sensitivity branch
sensitivity_branch_fingerprint(specification)sensitivity_branch_fingerprint(specification)
specification |
One-row specification table or named list. |
An R object containing stable fingerprint of a sensitivity branch. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Leverage is descriptive variation in mean effect across option levels; it is not causal attribution of researcher decisions.
sensitivity_decision_leverage(x, effect = "effect")sensitivity_decision_leverage(x, effect = "effect")
x |
Sensitivity result. |
effect |
Effect column. |
A tabular R object containing decision leverage of each analytical choice; rows represent analysis units and columns contain the returned quantities.
Fragility index across analysis specifications
sensitivity_fragility_index(x, effect = "effect", threshold = 0)sensitivity_fragility_index(x, effect = "effect", threshold = 0)
x |
Sensitivity result. |
effect |
Effect column. |
threshold |
Decision threshold. |
A numeric value or vector containing fragility index across analysis specifications.
Machine-readable multiverse manifest
sensitivity_multiverse_manifest(x)sensitivity_multiverse_manifest(x)
x |
Sensitivity result. |
A data frame containing machine-readable multiverse manifest. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Rank stability across specifications
sensitivity_rank_stability(x, id = NULL, rank = NULL, specification = NULL)sensitivity_rank_stability(x, id = NULL, rank = NULL, specification = NULL)
x |
Data frame or list of ranking vectors. |
id |
Optional item identifier when x is a long data frame. |
rank |
Optional rank/value column when x is a long data frame. |
specification |
Optional specification column. |
An R object containing rank stability across specifications. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Effect-sign stability across specifications
sensitivity_sign_stability(x, effect = "effect")sensitivity_sign_stability(x, effect = "effect")
x |
Sensitivity result. |
effect |
Effect column. |
A numeric value or vector containing effect-sign stability across specifications.
Significance-decision stability across specifications
sensitivity_significance_stability(x, p_value = "p_value", alpha = 0.05)sensitivity_significance_stability(x, p_value = "p_value", alpha = 0.05)
x |
Sensitivity result. |
p_value |
P-value column. |
alpha |
Decision threshold. |
A numeric value or vector containing significance-decision stability across specifications.
Substantive-threshold stability across specifications
sensitivity_threshold_stability( x, effect = "effect", threshold = 0, direction = c("above", "below", "absolute") )sensitivity_threshold_stability( x, effect = "effect", threshold = 0, direction = c("above", "below", "absolute") )
x |
Sensitivity result. |
effect |
Effect column. |
threshold |
Threshold. |
direction |
'above', 'below', or 'absolute'. |
A numeric value or vector containing substantive-threshold stability across specifications.
Convert scanpaths to process/sequence package contracts
as_procdata_sequence( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE ) as_traminer_sequence( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE, create_object = FALSE ) as_seqhmm_data( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE )as_procdata_sequence( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE ) as_traminer_sequence( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE, create_object = FALSE ) as_seqhmm_data( x, source = c("visits", "fixations", "samples"), collapse_consecutive = TRUE )
x |
An 'eye_dataset'. |
source |
AOI sequence source. |
collapse_consecutive |
Whether to collapse repeated adjacent states. |
create_object |
For 'as_traminer_sequence()', whether to return a native 'TraMineR' sequence object instead of the package-neutral wide table. |
A package-compatible representation.
Extract session facet effects
session_facet_effects(object, channel = c("response", "process"))session_facet_effects(object, channel = c("response", "process"))
object |
A fitted eyeprocess model or audit object. |
channel |
Measurement channel to inspect. |
An object of class "eye_process_facet_effects", stored as a named list, with components "facet", "column", "channel", "random_effects", "variance_component". It contains session facet effects and associated metadata or diagnostics needed to interpret the result.
Generates accuracy, response time, process features, theory-defined strategy, dynamic AOI states, pupil trajectories, measurement error, missing process data, DIF, and local dependence for validation studies.
simulate_advanced_process_data( n_person = 100L, n_item = 20L, n_time = 30L, n_states = 3L, ability_speed_correlation = -0.3, gaze_effect = 0.35, feature_reliability = 0.7, missing_process = 0, state_misclassification = 0, pupil_ar1 = 0.6, luminance_effect = 0, dif_effect = 0, local_dependence = 0, seed = 1L )simulate_advanced_process_data( n_person = 100L, n_item = 20L, n_time = 30L, n_states = 3L, ability_speed_correlation = -0.3, gaze_effect = 0.35, feature_reliability = 0.7, missing_process = 0, state_misclassification = 0, pupil_ar1 = 0.6, luminance_effect = 0, dif_effect = 0, local_dependence = 0, seed = 1L )
n_person |
Number of persons. |
n_item |
Number of items. |
n_time |
Pupil time bins. |
n_states |
Number of AOI states. |
ability_speed_correlation |
Correlation between ability and speed. |
gaze_effect |
Process-feature coefficient in the response model. |
feature_reliability |
Reliability of observed gaze features. |
missing_process |
Fraction of process observations set missing. |
state_misclassification |
Probability of AOI-state misclassification. |
pupil_ar1 |
AR(1) coefficient for pupil noise. |
luminance_effect |
Effect of simulated luminance on pupil size. |
dif_effect |
Logit-scale DIF effect for the focal group on flagged items. |
local_dependence |
Shared testlet-effect standard deviation. |
seed |
Random seed. |
A list with trial data, state data, pupil data, and truth.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
simulate_dynamic_irtree_data(n_person = 100L, n_item = 20L, transitions_per_trial = 8L, states = c("prompt", "evidence", "options"), beta_response = 0.5, person_sd = 0.4, item_sd = 0.3, irregular_time = TRUE, state_misclassification = 0, missing_state = 0, structural_zeros = NULL, seed = 1L)simulate_dynamic_irtree_data(n_person = 100L, n_item = 20L, transitions_per_trial = 8L, states = c("prompt", "evidence", "options"), beta_response = 0.5, person_sd = 0.4, item_sd = 0.3, irregular_time = TRUE, state_misclassification = 0, missing_state = 0, structural_zeros = NULL, seed = 1L)
n_person |
Number of persons and items. |
n_item |
Number of persons and items. |
transitions_per_trial |
Number of transitions per person-item trial. |
states |
State labels. |
beta_response |
Response effect on transitions. |
person_sd |
Person/item heterogeneity. |
item_sd |
Person/item heterogeneity. |
irregular_time |
Whether to generate irregular time gaps. |
state_misclassification |
Destination-state error probability. |
missing_state |
Missing destination-state probability. |
structural_zeros |
Optional forbidden transitions. |
seed |
Random seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Run a catR adaptive-testing simulation without fallback substitution
simulate_eyeprocess_catr(itemBank, trueTheta = 0, ..., engine = "catR")simulate_eyeprocess_catr(itemBank, trueTheta = 0, ..., engine = "catR")
itemBank |
Item bank supplied to catR. |
trueTheta |
Known true latent-trait value used for CAT simulation. |
... |
Additional arguments passed to the selected method or external engine. |
engine |
Requested estimation or analysis engine. |
An object of class "eye_external_irt_fit", stored as a named list, with components "status", "engine", "fit", "call". It contains a catR adaptive-testing simulation without fallback substitution and associated metadata or diagnostics needed to interpret the result.
Simulate dichotomous IRT responses with optional local dependence and missingness
simulate_eyeprocess_irt_binary( n_persons = 500L, items, theta = NULL, missing_rate = 0, testlet_sd = 0, seed = 1L, D = 1 )simulate_eyeprocess_irt_binary( n_persons = 500L, items, theta = NULL, missing_rate = 0, testlet_sd = 0, seed = 1L, D = 1 )
n_persons |
Number of persons. |
items |
Item-parameter data frame or item collection. |
theta |
Latent-trait value or vector of latent-trait values. |
missing_rate |
Proportion of responses or observations set missing. |
testlet_sd |
Standard deviation of simulated testlet effects. |
seed |
Random-number seed for reproducible execution. |
D |
Logistic scaling constant. |
An object of class "eye_irt_simulation", stored as a named list, with components "responses", "probabilities", "theta", "items", "missing_rate", "testlet_sd", "seed". It contains dichotomous IRT responses with optional local dependence and missingness and associated metadata or diagnostics needed to interpret the result.
Simulate data from a model or registered model specification
simulate_from_model(model, ...)simulate_from_model(model, ...)
model |
Registered model id/specification, simulation function, or an object exposing a 'simulate_fun' function. |
... |
Arguments passed to the simulator. |
An R object containing data from a model or registered model specification. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
simulate_gaze_diffusion_data(n_person = 80L, n_item = 20L, trials_per_item = 1L, gaze_effect = 0.35, contaminant_fraction = 0.02, time_step = 0.002, max_decision_time = 10, seed = 1L)simulate_gaze_diffusion_data(n_person = 80L, n_item = 20L, trials_per_item = 1L, gaze_effect = 0.35, contaminant_fraction = 0.02, time_step = 0.002, max_decision_time = 10, seed = 1L)
n_person |
Number of participants/items. |
n_item |
Number of participants/items. |
trials_per_item |
Replications per person-item. |
gaze_effect |
Drift effect of gaze feature. |
contaminant_fraction |
Contaminant fraction. |
time_step |
Time step for the built-in Wiener discretization fallback. |
max_decision_time |
Maximum fallback decision time in seconds. |
seed |
Seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Simulate from a registered multimodal IRT model
simulate_irt_model(spec, ..., allow_experimental = TRUE)simulate_irt_model(spec, ..., allow_experimental = TRUE)
spec |
IRT model or validation specification. |
... |
Additional arguments passed to the selected model, engine, or method. |
allow_experimental |
Whether experimental models are permitted. |
An R object containing from a registered multimodal IRT model. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Generates deterministic synthetic person-item observations for staged M0-M3 development. Gaze is a non-negative count process and pupil is a neutral pupil-responsivity channel with explicit nuisance effects.
simulate_multimodal_irt( n_person = 120L, n_item = 20L, seed = 42L, latent_cor = diag(4L), rt_sd = 0.3, gaze_size = 8, pupil_sd = 0.2, pupil_luminance = -0.2, gaze_x_effect = 0.08, gaze_y_effect = -0.05, missing_fraction = 0 )simulate_multimodal_irt( n_person = 120L, n_item = 20L, seed = 42L, latent_cor = diag(4L), rt_sd = 0.3, gaze_size = 8, pupil_sd = 0.2, pupil_luminance = -0.2, gaze_x_effect = 0.08, gaze_y_effect = -0.05, missing_fraction = 0 )
n_person, n_item
|
Positive integers. |
seed |
Random seed. |
latent_cor |
4x4 correlation matrix for ability, speed, gaze process, and pupil responsivity. |
rt_sd |
Residual log-RT SD. |
gaze_size |
Negative-binomial size. |
pupil_sd |
Residual pupil-channel SD. |
pupil_luminance, gaze_x_effect, gaze_y_effect
|
Nuisance coefficients. |
missing_fraction |
Independent channel dropout fraction used for a baseline stress condition. |
An 'eye_multimodal_simulation'.
Simulates the same level-1 likelihood used by [fit_multimodal_m2()] and retains all person/item hyperparameters and realized latent parameters as truth. Channel-specific dropout is applied only after the complete generative data have been created.
simulate_multimodal_m2( n_person = 100L, n_item = 10L, mu_item = c(difficulty = 0, time_intensity = 4, gaze_intensity = 3.5), sd_person = c(ability = 1, speed = 0.5, gaze_process = 0.5), cor_person = matrix(c(1, 0.3, -0.3, 0.3, 1, -0.25, -0.3, -0.25, 1), 3L, 3L, byrow = TRUE), sd_item = c(difficulty = 0.75, time_intensity = 0.35, gaze_intensity = 0.6), cor_item = matrix(c(1, 0.25, 0.2, 0.25, 1, 0.3, 0.2, 0.3, 1), 3L, 3L, byrow = TRUE), nu_range = c(0.5, 0.8), gaze_shape = c(shape = 2, scale = 6), dropout = c(response = 0, rt = 0, gaze = 0), seed = 20260814L )simulate_multimodal_m2( n_person = 100L, n_item = 10L, mu_item = c(difficulty = 0, time_intensity = 4, gaze_intensity = 3.5), sd_person = c(ability = 1, speed = 0.5, gaze_process = 0.5), cor_person = matrix(c(1, 0.3, -0.3, 0.3, 1, -0.25, -0.3, -0.25, 1), 3L, 3L, byrow = TRUE), sd_item = c(difficulty = 0.75, time_intensity = 0.35, gaze_intensity = 0.6), cor_item = matrix(c(1, 0.25, 0.2, 0.25, 1, 0.3, 0.2, 0.3, 1), 3L, 3L, byrow = TRUE), nu_range = c(0.5, 0.8), gaze_shape = c(shape = 2, scale = 6), dropout = c(response = 0, rt = 0, gaze = 0), seed = 20260814L )
n_person |
Number of persons. |
n_item |
Number of items. |
mu_item |
Means for item difficulty, log-time intensity, and log-gaze intensity. |
sd_person |
Person-side standard deviations for ability, speed, and gaze-process propensity. |
cor_person |
Person-side correlation matrix. |
sd_item |
Item-side standard deviations. |
cor_item |
Item-side correlation matrix. |
nu_range |
Uniform range for item time-discrimination parameters. |
gaze_shape |
Shape/scale parameters for inverse-gamma generation of negative-binomial shape parameters. |
dropout |
Named probabilities for response, RT, and gaze missingness. |
seed |
Reproducibility seed. |
An 'eye_multimodal_m2_simulation'.
Generates a four-dimensional correlated person process and four-dimensional correlated item process, explicit pupil nuisance variables, complete-data truth, device/session metadata, blink/interpolation indicators and pupil dropout. Pupil scenarios include informative, weak, null, redundant and confound-only conditions so that validation includes cases where pupil should add no defensible psychometric information.
simulate_multimodal_m3( n_person = 120L, n_item = 12L, pupil_signal = c("informative", "weak", "null", "redundant", "confounded"), pupil_missingness = c("mcar", "quality", "gaze", "ability", "device", "none"), mu_item = c(difficulty = 0, time_intensity = 4, gaze_intensity = 3.5, pupil_intensity = 0), sd_person = c(ability = 1, speed = 0.5, gaze_process = 0.5, pupil_responsivity = 0.55), cor_person = matrix(c(1, 0.3, -0.3, 0.2, 0.3, 1, -0.25, -0.15, -0.3, -0.25, 1, 0.25, 0.2, -0.15, 0.25, 1), 4L, 4L, byrow = TRUE), sd_item = c(difficulty = 0.75, time_intensity = 0.35, gaze_intensity = 0.6, pupil_intensity = 0.4), cor_item = matrix(c(1, 0.25, 0.2, 0.1, 0.25, 1, 0.3, 0.15, 0.2, 0.3, 1, 0.2, 0.1, 0.15, 0.2, 1), 4L, 4L, byrow = TRUE), nu_range = c(0.5, 0.8), gaze_shape = c(shape = 2, scale = 6), pupil_noise = 0.65, confound_strength = c(baseline = 0.25, luminance = -0.35, gaze_x = 0.12, gaze_y = -0.1, quality = 0.2, blink = -0.18, interpolated = -0.12, time_on_task = 0.15), dropout = c(response = 0, rt = 0, gaze = 0.05, pupil = 0.12), device_effect = 0, session_effect = 0, seed = 20260815L )simulate_multimodal_m3( n_person = 120L, n_item = 12L, pupil_signal = c("informative", "weak", "null", "redundant", "confounded"), pupil_missingness = c("mcar", "quality", "gaze", "ability", "device", "none"), mu_item = c(difficulty = 0, time_intensity = 4, gaze_intensity = 3.5, pupil_intensity = 0), sd_person = c(ability = 1, speed = 0.5, gaze_process = 0.5, pupil_responsivity = 0.55), cor_person = matrix(c(1, 0.3, -0.3, 0.2, 0.3, 1, -0.25, -0.15, -0.3, -0.25, 1, 0.25, 0.2, -0.15, 0.25, 1), 4L, 4L, byrow = TRUE), sd_item = c(difficulty = 0.75, time_intensity = 0.35, gaze_intensity = 0.6, pupil_intensity = 0.4), cor_item = matrix(c(1, 0.25, 0.2, 0.1, 0.25, 1, 0.3, 0.15, 0.2, 0.3, 1, 0.2, 0.1, 0.15, 0.2, 1), 4L, 4L, byrow = TRUE), nu_range = c(0.5, 0.8), gaze_shape = c(shape = 2, scale = 6), pupil_noise = 0.65, confound_strength = c(baseline = 0.25, luminance = -0.35, gaze_x = 0.12, gaze_y = -0.1, quality = 0.2, blink = -0.18, interpolated = -0.12, time_on_task = 0.15), dropout = c(response = 0, rt = 0, gaze = 0.05, pupil = 0.12), device_effect = 0, session_effect = 0, seed = 20260815L )
n_person, n_item
|
Design size. |
pupil_signal |
Pupil signal scenario. |
pupil_missingness |
Missingness stress mechanism. |
mu_item |
Named numeric vector of population means for item difficulty, response-time intensity, gaze intensity, and pupil intensity. |
sd_person |
Named positive numeric vector of population standard deviations for person ability, speed, gaze-process propensity, and pupil responsivity. |
cor_person |
A 4 x 4 correlation matrix for person ability, speed, gaze-process, and pupil-responsivity effects, in that order. |
sd_item |
Named positive numeric vector of population standard deviations for item difficulty, response-time intensity, gaze intensity, and pupil intensity. |
cor_item |
A 4 x 4 correlation matrix for item difficulty, response-time intensity, gaze intensity, and pupil intensity, in that order. |
nu_range |
Length-two positive increasing numeric vector giving the lower and upper bounds for the item-specific response-time inverse-scale parameter 'nu'; the log-response-time residual standard deviation is '1 / nu'. |
gaze_shape |
Named positive numeric vector with elements 'shape' and 'scale' defining the inverse-gamma generator for item-specific negative-binomial gaze dispersion. |
pupil_noise |
Residual SD for the pupil channel. |
confound_strength |
Named standardized nuisance coefficients. |
dropout |
Named base dropout probabilities for response, RT, gaze, pupil. |
device_effect, session_effect
|
Additive pupil measurement shifts. |
seed |
Reproducibility seed. |
An 'eye_multimodal_m3_simulation' retaining complete truth.
Generates deterministic synthetic response, RT, gaze, pupil, nuisance, and ordered latent-state data for software validation and methodological stress testing. Synthetic state labels are known truth for recovery only and are not psychological constructs.
simulate_multimodal_m4( n_person = 80L, n_item = 12L, n_session = 1L, n_states = 2L, scenario = c("clear", "weak", "null", "persistent", "rapid_switch", "trait_conditioned", "rt_redundant", "gaze_redundant", "pupil_redundant", "nuisance_confounded", "device_confounded"), missingness = c("none", "mcar", "quality", "gaze", "pupil_quality", "device", "state_dependent"), missing_rate = 0.08, seed = 20260820L )simulate_multimodal_m4( n_person = 80L, n_item = 12L, n_session = 1L, n_states = 2L, scenario = c("clear", "weak", "null", "persistent", "rapid_switch", "trait_conditioned", "rt_redundant", "gaze_redundant", "pupil_redundant", "nuisance_confounded", "device_confounded"), missingness = c("none", "mcar", "quality", "gaze", "pupil_quality", "device", "state_dependent"), missing_rate = 0.08, seed = 20260820L )
n_person, n_item
|
Number of persons and total item trials per person. |
n_session |
Number of non-overlapping ordered sessions per person. |
n_states |
True latent-state count. May be 1 through 4. |
scenario |
State/data-generating scenario: 'clear', 'weak', 'null', 'persistent', 'rapid_switch', 'trait_conditioned', 'rt_redundant', 'gaze_redundant', 'pupil_redundant', 'nuisance_confounded', or 'device_confounded'. |
missingness |
Missingness stress mechanism. |
missing_rate |
Base channel-missingness probability. |
seed |
Deterministic random seed. |
An 'eye_multimodal_m4_simulation' containing 'data', full generating 'truth', scenario metadata, and interpretation boundary.
Simulate pre-registered presentation variants for review
simulate_presentation_variants( audit, line_spacing_multiplier = 1.25, key_term_highlighting = TRUE )simulate_presentation_variants( audit, line_spacing_multiplier = 1.25, key_term_highlighting = TRUE )
audit |
An accessibility audit. |
line_spacing_multiplier |
Example line-spacing multiplier for flagged rows. |
key_term_highlighting |
Whether the simulated review variant highlights key terms. |
An R object containing pre-registered presentation variants for review. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
This is a design simulator, not a production testing engine.
simulate_process_cat( item_bank, true_theta = 0, n_items = 10L, weights = c(response = 1, rt = 0, process = 0), burden_weight = 0, seed = 1 )simulate_process_cat( item_bank, true_theta = 0, n_items = 10L, weights = c(response = 1, rt = 0, process = 0), burden_weight = 0, seed = 1 )
item_bank |
Value supplied to 'item_bank'; see Details for its model-specific role. |
true_theta |
Simulated true latent-trait value or values. |
n_items |
Number of items. |
weights |
Weights used to combine information components. |
burden_weight |
Penalty applied to expected burden. |
seed |
Random-number seed. |
An object of class "eye_process_cat_simulation", "data.frame", stored as a data frame, containing a simple process-aware CAT policy and associated metadata needed to interpret the result.
This simulator is a neutral software-validation fixture, not a substantive psychological data-generating model. The generated process channels should not be interpreted as mental-state measurements.
simulate_process_validation_data( condition, replication = 1L, seed = NULL, beta = 0.35 )simulate_process_validation_data( condition, replication = 1L, seed = NULL, beta = 0.35 )
condition |
One-row validation condition. |
replication |
Replication index. |
seed |
Optional seed override. |
beta |
Known effect of 'x' on the generic process outcome. |
List with 'data' and 'truth'.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
simulate_strategy_mixture_data(n_person = 100L, n_item = 20L, signatures, trials_per_item = 1L, strategy_prevalence = NULL, feature_sd = 0.6, seed = 1L)simulate_strategy_mixture_data(n_person = 100L, n_item = 20L, signatures, trials_per_item = 1L, strategy_prevalence = NULL, feature_sd = 0.6, seed = 1L)
n_person |
Number of participants. |
n_item |
Number of items. |
signatures |
Strategy signature matrix. |
trials_per_item |
Trials per person-item combination. |
strategy_prevalence |
Strategy prevalence. |
feature_sd |
Feature residual SD. |
seed |
Seed. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Run simulation-based calibration
simulation_based_calibration( simulator, fitter, posterior_draws, truth_extractor, replications = 100L, seed = 1L, ... )simulation_based_calibration( simulator, fitter, posterior_draws, truth_extractor, replications = 100L, seed = 1L, ... )
simulator |
Function returning simulated data and named truth. |
fitter |
Function receiving one simulation result. |
posterior_draws |
Function returning a numeric matrix/data frame whose columns are named parameters. |
truth_extractor |
Function returning a named numeric truth vector. |
replications |
Number of simulated data sets. |
seed |
Random seed. |
... |
Passed to 'simulator()'. |
An 'eye_sbc' object with parameter ranks and calibration summaries.
Compute a simulation-based calibration rank statistic
simulation_rank_statistic(truth, draws, seed = NULL)simulation_rank_statistic(truth, draws, seed = NULL)
truth |
Scalar simulated truth. |
draws |
Posterior draws for the same parameter. |
seed |
Seed used only to randomize ties. |
A numeric value or vector containing a simulation-based calibration rank statistic.
Create or normalize a software-paper claim matrix
software_paper_claim_matrix( claim, evidence_id = NA_character_, evidence_type = NA_character_, status = "pending", scope = NA_character_, source = NA_character_ )software_paper_claim_matrix( claim, evidence_id = NA_character_, evidence_type = NA_character_, status = "pending", scope = NA_character_, source = NA_character_ )
claim |
Claim text. |
evidence_id |
Evidence identifiers. |
evidence_type |
Evidence type. |
status |
Status such as supported, qualified, pending, or unsupported. |
scope |
Explicit scope/qualification. |
source |
Optional source/location. |
A data frame containing or normalize a software-paper claim matrix. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Compute descriptive evidence coverage
software_paper_coverage(x, supported = c("supported", "qualified"))software_paper_coverage(x, supported = c("supported", "qualified"))
x |
Evidence bundle or claim matrix. |
supported |
Status labels counted as covered. |
A data frame containing descriptive evidence coverage. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Construct a software-paper evidence bundle
software_paper_evidence_bundle( claims = NULL, validation = NULL, examples = NULL, articles = NULL, benchmarks = NULL, reproducibility = NULL, metadata = list() )software_paper_evidence_bundle( claims = NULL, validation = NULL, examples = NULL, articles = NULL, benchmarks = NULL, reproducibility = NULL, metadata = list() )
claims |
Claim table or list. |
validation |
Validation evidence table/list. |
examples |
Optional example inventory. |
articles |
Optional article inventory. |
benchmarks |
Optional benchmark evidence. |
reproducibility |
Optional reproducibility fingerprint. |
metadata |
Optional metadata. |
A named list with components "schema_version", "claims", "validation", "examples", "articles", "benchmarks", "reproducibility", "metadata", "created_utc", containing a software-paper evidence bundle and associated metadata or diagnostics.
Identify gaps in a software-paper evidence bundle
software_paper_gap_analysis(x)software_paper_gap_analysis(x)
x |
Evidence bundle. |
A named list with components "requirement_gaps", "claim_gaps", containing gaps in a software-paper evidence bundle and associated metadata or diagnostics.
Readiness is defined only against caller-specified requirements; it is not a journal acceptance prediction.
software_paper_readiness( x, required_statuses = c("supported", "qualified"), require_validation = TRUE, require_reproducibility = TRUE, require_examples = TRUE, require_articles = TRUE )software_paper_readiness( x, required_statuses = c("supported", "qualified"), require_validation = TRUE, require_reproducibility = TRUE, require_examples = TRUE, require_articles = TRUE )
x |
Evidence bundle. |
required_statuses |
Statuses allowed for claims. |
require_validation |
Require non-empty validation evidence. |
require_reproducibility |
Require a reproducibility fingerprint. |
require_examples |
Require examples. |
require_articles |
Require articles. |
A named list with components "ready", "checks", "interpretation", containing descriptive software-paper readiness audit and associated metadata or diagnostics.
Summarise validation evidence for a software paper
software_paper_validation_table(x)software_paper_validation_table(x)
x |
Validation result/evidence matrix/table. |
An R object containing validation evidence for a software paper. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Fraction of planned specifications successfully evaluated
specification_coverage(x)specification_coverage(x)
x |
Sensitivity result. |
A numeric value or vector containing fraction of planned specifications successfully evaluated.
Prepare ordered specification-curve data
specification_curve_data(x, effect = "effect", lower = NULL, upper = NULL)specification_curve_data(x, effect = "effect", lower = NULL, upper = NULL)
x |
Sensitivity result. |
effect |
Effect column. |
lower |
Optional lower interval column. |
upper |
Optional upper interval column. |
A tabular R object containing ordered specification-curve data; rows represent analysis units and columns contain the returned quantities.
Split-half reliability for a trial-level process measure
split_half_process_reliability( data, person, trial, measure, split = c("odd_even", "random"), repetitions = 100L, seed = 1L, aggregate_fun = mean )split_half_process_reliability( data, person, trial, measure, split = c("odd_even", "random"), repetitions = 100L, seed = 1L, aggregate_fun = mean )
data |
Long trial-level data. |
person |
Participant column. |
trial |
Trial column. |
measure |
Measure column. |
split |
Odd/even or repeated random split. |
repetitions |
Number of random splits. |
seed |
Seed. |
aggregate_fun |
Within-half aggregation function. |
A tabular R object containing split-half reliability for a trial-level process measure; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
split_validation_plan(plan, chunks = NULL)split_validation_plan(plan, chunks = NULL)
plan |
Validation plan. |
chunks |
Optional chunk identifiers. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compute and report separable eye-tracking quality dimensions: target-referenced accuracy, RMS sample-to-sample precision, spatial standard-deviation precision, BCEA, effective sampling behavior, sampling jitter, valid-data fraction, and data loss. User thresholds create review flags only and never trigger automatic exclusion.
validate_gaze_quality_inputs(data, x = "gaze_x", y = "gaze_y", time = NULL, target_x = NULL, target_y = NULL, by = NULL, unit = "degrees", time_unit = "ms", unit_column = NULL) compute_gaze_accuracy(data, x = "gaze_x", y = "gaze_y", target_x = "target_x", target_y = "target_y", by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL) compute_rms_s2s(data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = c("2d", "horizontal", "vertical"), time_unit = "ms", max_gap_ms = NULL) compute_gaze_sd_precision(data, x = "gaze_x", y = "gaze_y", by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL) compute_bcea(data, x = "gaze_x", y = "gaze_y", by = NULL, probability = 0.68, unit = "degrees", output_unit = NULL, geometry = NULL) compute_gaze_precision(data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL, probability = 0.68, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = "2d", time_unit = "ms", max_gap_ms = NULL) estimate_sampling_interval(data, time = "timestamp_ms", by = NULL, time_unit = "ms") estimate_sampling_jitter(data, time = "timestamp_ms", by = NULL, time_unit = "ms") estimate_effective_sampling_rate(data, time = "timestamp_ms", by = NULL, time_unit = "ms", nominal_sampling_hz = NULL, dropped_interval_factor = 1.5, x = NULL, y = NULL, valid = NULL) compute_valid_sample_fraction(data, x = "gaze_x", y = "gaze_y", valid = NULL, by = NULL) compute_gaze_data_loss(data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", valid = NULL, missing_reason = NULL, by = NULL, time_unit = "ms") summarise_spatial_quality(data, x = "gaze_x", y = "gaze_y", target_x = "target_x", target_y = "target_y", time = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = "2d", time_unit = "ms", max_gap_ms = NULL, probability = 0.68) summarise_sampling_quality(data, time = "timestamp_ms", by = NULL, time_unit = "ms", nominal_sampling_hz = NULL, dropped_interval_factor = 1.5, x = NULL, y = NULL, valid = NULL) create_gaze_quality_report(data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", target_x = "target_x", target_y = "target_y", valid = NULL, missing_reason = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, time_unit = "ms", nominal_sampling_hz = NULL, bcea_probability = 0.68, max_gap_ms = NULL, thresholds = NULL, preprocessing_spec = NULL, event_detector = NULL, aoi_specification = NULL, quality_rules = NULL, model_specification = NULL, software_version = NULL, unit_column = NULL) compare_gaze_quality_sessions(data, session = "session_id", by = NULL, ...) compare_gaze_quality_conditions(data, condition = "condition", by = NULL, ...) plot_gaze_accuracy(report, metric = "accuracy_mean", ...) plot_gaze_precision(report, metric = "precision_rms_s2s", ...) plot_bcea(report, ...) plot_sampling_intervals(data, time = "timestamp_ms", time_unit = "ms", ...) plot_gaze_quality_dashboard(report, ...) report_gaze_quality(report, digits = 3L) simulate_gaze_quality_calibration(seed = 20260918L, samples_per_target = 18L, nominal_sampling_hz = 60)validate_gaze_quality_inputs(data, x = "gaze_x", y = "gaze_y", time = NULL, target_x = NULL, target_y = NULL, by = NULL, unit = "degrees", time_unit = "ms", unit_column = NULL) compute_gaze_accuracy(data, x = "gaze_x", y = "gaze_y", target_x = "target_x", target_y = "target_y", by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL) compute_rms_s2s(data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = c("2d", "horizontal", "vertical"), time_unit = "ms", max_gap_ms = NULL) compute_gaze_sd_precision(data, x = "gaze_x", y = "gaze_y", by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL) compute_bcea(data, x = "gaze_x", y = "gaze_y", by = NULL, probability = 0.68, unit = "degrees", output_unit = NULL, geometry = NULL) compute_gaze_precision(data, x = "gaze_x", y = "gaze_y", time = NULL, by = NULL, probability = 0.68, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = "2d", time_unit = "ms", max_gap_ms = NULL) estimate_sampling_interval(data, time = "timestamp_ms", by = NULL, time_unit = "ms") estimate_sampling_jitter(data, time = "timestamp_ms", by = NULL, time_unit = "ms") estimate_effective_sampling_rate(data, time = "timestamp_ms", by = NULL, time_unit = "ms", nominal_sampling_hz = NULL, dropped_interval_factor = 1.5, x = NULL, y = NULL, valid = NULL) compute_valid_sample_fraction(data, x = "gaze_x", y = "gaze_y", valid = NULL, by = NULL) compute_gaze_data_loss(data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", valid = NULL, missing_reason = NULL, by = NULL, time_unit = "ms") summarise_spatial_quality(data, x = "gaze_x", y = "gaze_y", target_x = "target_x", target_y = "target_y", time = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, dimension = "2d", time_unit = "ms", max_gap_ms = NULL, probability = 0.68) summarise_sampling_quality(data, time = "timestamp_ms", by = NULL, time_unit = "ms", nominal_sampling_hz = NULL, dropped_interval_factor = 1.5, x = NULL, y = NULL, valid = NULL) create_gaze_quality_report(data, x = "gaze_x", y = "gaze_y", time = "timestamp_ms", target_x = "target_x", target_y = "target_y", valid = NULL, missing_reason = NULL, by = NULL, unit = "degrees", output_unit = NULL, geometry = NULL, time_unit = "ms", nominal_sampling_hz = NULL, bcea_probability = 0.68, max_gap_ms = NULL, thresholds = NULL, preprocessing_spec = NULL, event_detector = NULL, aoi_specification = NULL, quality_rules = NULL, model_specification = NULL, software_version = NULL, unit_column = NULL) compare_gaze_quality_sessions(data, session = "session_id", by = NULL, ...) compare_gaze_quality_conditions(data, condition = "condition", by = NULL, ...) plot_gaze_accuracy(report, metric = "accuracy_mean", ...) plot_gaze_precision(report, metric = "precision_rms_s2s", ...) plot_bcea(report, ...) plot_sampling_intervals(data, time = "timestamp_ms", time_unit = "ms", ...) plot_gaze_quality_dashboard(report, ...) report_gaze_quality(report, digits = 3L) simulate_gaze_quality_calibration(seed = 20260918L, samples_per_target = 18L, nominal_sampling_hz = 60)
data |
A data frame or object coercible to a data frame. |
x, y
|
Columns containing horizontal and vertical gaze coordinates. |
time |
Timestamp column. |
target_x, target_y
|
Known validation-target coordinate columns. |
by |
Optional grouping columns defining the quality-analysis unit. |
unit |
Input coordinate unit: |
output_unit |
Optional explicit output coordinate unit. No conversion occurs when omitted. |
geometry |
Named screen/viewing geometry used only for explicit coordinate conversion. |
time_unit |
Timestamp unit: seconds, milliseconds, microseconds, or nanoseconds. |
unit_column |
Optional column used to reject mixed or conflicting coordinate units. For the canonical report, an existing |
dimension |
RMS-S2S dimension: two-dimensional, horizontal, or vertical. |
max_gap_ms |
Optional explicit maximum interval permitted for an adjacent-sample RMS pair. |
probability, bcea_probability
|
Probability region used for BCEA. |
nominal_sampling_hz |
Optional nominal hardware rate used for comparison and dropped-interval diagnostics. |
dropped_interval_factor |
Multiple of the nominal interval used to count long intervals. |
valid |
Optional explicit validity column. Missing gaze coordinates remain unavailable regardless of this flag. |
missing_reason |
Optional loss-reason column such as blink, tracker invalidity, or off-screen. |
thresholds |
Optional named study-specific review thresholds; these never exclude data automatically. |
preprocessing_spec, event_detector, aoi_specification, quality_rules, model_specification, software_version
|
Optional provenance fields preserved on the canonical report. |
session, condition
|
Grouping column used for convenience comparisons. |
report |
A quality-report-like data frame. |
metric |
Metric column plotted by a focused quality plot. |
digits |
Number of digits in the compact reporting string. |
seed |
Deterministic synthetic-data seed. |
samples_per_target |
Synthetic samples generated at each of nine validation targets. |
... |
Additional arguments forwarded to the canonical report or base plotting function. |
Accuracy is target-referenced error. RMS-S2S precision quantifies adjacent-sample
fluctuation and never bridges a missing sample. Spatial SD and BCEA quantify
point-cloud spread around a stable target and use population standard deviations
(denominator ). BCEA always records its probability level.
Effective sampling frequency is estimated from the effective sample count divided
by an estimated recording duration (timestamp span plus one median positive
inter-sample interval). When gaze coordinates and validity are supplied, the
effective sample count contains valid gaze observations with finite timestamps;
otherwise the metric describes the timestamp stream. With a nominal rate,
long_interval_count counts intervals exceeding the declared nominal-gap
multiple, whereas dropped_interval_count estimates the number of missing
nominal samples represented by those long intervals.
The package deliberately does not impose universal acceptability cutoffs. A
gaze_quality_report stores review flags and provenance while preserving all
analysis units for downstream filtering, weighting, stratification, sensitivity
analysis, and transparent reporting.
Metric functions return data frames with explicit units and diagnostics.
create_gaze_quality_report() returns a data frame of class
gaze_quality_report with a gaze_quality_provenance attribute.
Plot functions draw base-R diagnostics and invisibly return the plotted data.
report_gaze_quality() returns a compact character summary.
simulate_gaze_quality_calibration() returns a deterministic synthetic
nine-point validation data set spanning six quality profiles.
Dunn, M. J., et al. (2024). A minimal reporting guideline for eye-tracking studies. Behavior Research Methods.
Niehorster, D. C., et al. (2026). How to determine the quality of eye-tracking data: A tutorial. Behavior Research Methods. doi:10.3758/s13428-026-03039-4.
d <- simulate_gaze_quality_calibration(samples_per_target = 4) q <- create_gaze_quality_report( d, by = c("profile", "target_id"), valid = "valid", missing_reason = "missing_reason", nominal_sampling_hz = 60 ) head(q[c( "profile", "target_id", "accuracy_mean", "precision_rms_s2s", "bcea", "effective_sampling_hz", "data_loss_fraction" )]) report_gaze_quality(q)d <- simulate_gaze_quality_calibration(samples_per_target = 4) q <- create_gaze_quality_report( d, by = c("profile", "target_id"), valid = "valid", missing_reason = "missing_reason", nominal_sampling_hz = 60 ) head(q[c( "profile", "target_id", "accuracy_mean", "precision_rms_s2s", "bcea", "effective_sampling_hz", "data_loss_fraction" )]) report_gaze_quality(q)
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
storage_transaction_manifest(storage)storage_transaction_manifest(storage)
storage |
Storage object or path. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
strategy_aoi_sensitivity(datasets, spec, seed = 1L, ...)strategy_aoi_sensitivity(datasets, spec, seed = 1L, ...)
datasets |
Named list of alternative trial-level datasets. |
spec |
Strategy specification. |
seed |
Seed. |
... |
Fit arguments. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
strategy_classification_uncertainty(object, threshold = 0.70)strategy_classification_uncertainty(object, threshold = 0.70)
object |
Strategy fit. |
threshold |
Minimum modal probability. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
strategy_label_switching_diagnostics(object, tolerance = 1e-4)strategy_label_switching_diagnostics(object, tolerance = 1e-4)
object |
Strategy fit. |
tolerance |
Log-likelihood tolerance for equivalent starts. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
strategy_posterior_probabilities(object)strategy_posterior_probabilities(object)
object |
Strategy fit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Extract streaming score history
streaming_score_history(x)streaming_score_history(x)
x |
Object to process, inspect, compare, or plot. |
An R object containing streaming score history. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Stress test latent distribution
stress_test_latent_distribution(runner, replications = 50L, seed = 20260808L)stress_test_latent_distribution(runner, replications = 50L, seed = 20260808L)
runner |
Function that executes one stress-test scenario. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing stress test latent distribution and associated metadata needed to interpret the result.
Stress test local dependence
stress_test_local_dependence( runner, strengths = c(0, 0.2, 0.5, 0.8), replications = 50L, seed = 20260808L )stress_test_local_dependence( runner, strengths = c(0, 0.2, 0.5, 0.8), replications = 50L, seed = 20260808L )
runner |
Function that executes one stress-test scenario. |
strengths |
Local-dependence strengths to evaluate. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing stress test local dependence and associated metadata needed to interpret the result.
Stress test missingness
stress_test_missingness( runner, mechanisms = c("MCAR", "MAR", "MNAR_omission", "not_reached"), rates = c(0.05, 0.15, 0.3), replications = 50L, seed = 20260808L )stress_test_missingness( runner, mechanisms = c("MCAR", "MAR", "MNAR_omission", "not_reached"), rates = c(0.05, 0.15, 0.3), replications = 50L, seed = 20260808L )
runner |
Function that executes one stress-test scenario. |
mechanisms |
Missingness mechanisms to evaluate. |
rates |
Missingness rates to evaluate. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing stress test missingness and associated metadata needed to interpret the result.
Run a generic misspecification stress-test grid
stress_test_misspecification( scenarios, runner, replications = 50L, seed = 20260808L )stress_test_misspecification( scenarios, runner, replications = 50L, seed = 20260808L )
scenarios |
Data frame or named list describing scenarios. |
runner |
Function 'runner(scenario, replicate)' returning a one-row or tidy data frame. Errors are retained as classified failures. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing a generic misspecification stress-test grid and associated metadata needed to interpret the result.
Stress test preprocessing
stress_test_preprocessing( runner, variants, replications = 25L, seed = 20260808L )stress_test_preprocessing( runner, variants, replications = 25L, seed = 20260808L )
runner |
Function that executes one stress-test scenario. |
variants |
Preprocessing variants to evaluate. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing stress test preprocessing and associated metadata needed to interpret the result.
Stress-test an analysis under explicit synthetic corruptions
stress_test_process_pipeline( data, plans, analysis_fun, metric_fun = .ep09_default_sensitivity_extract, ... )stress_test_process_pipeline( data, plans, analysis_fun, metric_fun = .ep09_default_sensitivity_extract, ... )
data |
Baseline data. |
plans |
List of corruption plans. |
analysis_fun |
Function '(data, plan)'. |
metric_fun |
Function '(analysis_result, plan)' returning scalar/list/data.frame metrics. |
... |
Passed to 'apply_synthetic_corruption()'. |
An object of class "eye_process_stress_test", stored as a named list, with components "plans", "results", "baseline_hash", "created_at", "caveat". It contains stress-test an analysis under explicit synthetic corruptions and associated metadata or diagnostics needed to interpret the result.
Stress test speededness
stress_test_speededness( runner, proportions = c(0, 0.1, 0.25, 0.4), replications = 50L, seed = 20260808L )stress_test_speededness( runner, proportions = c(0, 0.1, 0.25, 0.4), replications = 50L, seed = 20260808L )
runner |
Function that executes one stress-test scenario. |
proportions |
Speededness proportions to evaluate. |
replications |
Number of simulation or validation replications. |
seed |
Random-number seed. |
An object of class "eye_irt_stress_test", "data.frame", stored as a data frame, containing stress test speededness and associated metadata needed to interpret the result.
Summarise stress-test metrics
stress_test_summary(x, metric = "effect")stress_test_summary(x, metric = "effect")
x |
Stress-test result. |
metric |
Numeric metric column. |
A data frame containing stress-test metrics. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Identify the empirical stress frontier for a metric
stress_tolerance_frontier(x, severity, metric, acceptable)stress_tolerance_frontier(x, severity, metric, acceptable)
x |
Stress-test result. |
severity |
Numeric corruption/severity column. |
metric |
Metric column. |
acceptable |
Function returning TRUE/FALSE for metric values. |
A data frame containing the empirical stress frontier for a metric. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
structural_transition_mask(states, forbidden = NULL, allowed = NULL, allow_self = TRUE, structural_zeros = NULL, allowed_transitions = NULL)structural_transition_mask(states, forbidden = NULL, allowed = NULL, allow_self = TRUE, structural_zeros = NULL, allowed_transitions = NULL)
states |
State labels. |
forbidden |
Two-column data frame/matrix of forbidden from-to pairs. |
allowed |
Two-column data frame/matrix of explicitly allowed pairs. |
allow_self |
Whether self transitions are allowed by default. |
structural_zeros |
Alias for 'forbidden'. |
allowed_transitions |
Alias for 'allowed'. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Summarise benchmark timing and memory by problem size
summarise_eye_benchmark(x)summarise_eye_benchmark(x)
x |
Benchmark result. |
A logical value or vector indicating benchmark timing and memory by problem size.
Summarise executed measurement-stress evidence
summarise_eyeprocess_stress_evidence(x)summarise_eyeprocess_stress_evidence(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A tabular R object containing executed measurement-stress evidence; rows represent analysis units and columns contain the returned quantities.
Summarise process negative controls
summarise_process_negative_controls(x, effect = "effect", threshold = 0)summarise_process_negative_controls(x, effect = "effect", threshold = 0)
x |
Negative-control result. |
effect |
Effect column. |
threshold |
Optional absolute effect threshold. |
A logical value or vector indicating process negative controls.
Summarise process sensitivity results
summarise_process_sensitivity( x, effect = "effect", p_value = NULL, threshold = 0, alpha = 0.05 )summarise_process_sensitivity( x, effect = "effect", p_value = NULL, threshold = 0, alpha = 0.05 )
x |
Sensitivity result. |
effect |
Effect column. |
p_value |
Optional p-value column. |
threshold |
Optional substantive effect threshold. |
alpha |
Significance threshold used only when 'p_value' is supplied. |
A data frame containing process sensitivity results. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Summarise a process-validation result
summarise_process_validation(x, by = NULL)summarise_process_validation(x, by = NULL)
x |
Validation result. |
by |
Optional grouping variables in addition to parameter. |
A tabular R object containing a process-validation result; rows represent analysis units and columns contain the returned quantities.
Summarise an acceptance matrix
summarise_validation_acceptance(x, by = character())summarise_validation_acceptance(x, by = character())
x |
Object to validate, summarize, verify, or otherwise process. |
by |
Grouping variables or aggregation level. |
A tabular R object containing an acceptance matrix; rows represent analysis units and columns contain the returned quantities.
Summarise parameter recovery
summarize_parameter_recovery( results, by = c("scenario", "engine", "parameter"), interval_level = 0.95 )summarize_parameter_recovery( results, by = c("scenario", "engine", "parameter"), interval_level = 0.95 )
results |
Canonical or raw recovery results. |
by |
Grouping columns. |
interval_level |
Nominal interval level, used only for labelling. |
A tabular R object containing parameter recovery; rows represent analysis units and columns contain the returned quantities.
Summarize extracted process windows
summarize_process_windows(x, by = NULL)summarize_process_windows(x, by = NULL)
x |
'eye_process_windows' object. |
by |
Optional grouping columns present in the extracted table. |
An R object containing extracted process windows. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Summarize a validation bundle object
## S3 method for class 'eye_validation_bundle' summary(object, ...)## S3 method for class 'eye_validation_bundle' summary(object, ...)
object |
Object supplied to the S3 method. |
... |
Additional arguments passed to the underlying method or helper. |
A named list with components "model_name", "manifest", "report", containing a validation bundle object and associated metadata or diagnostics.
S3 method supporting a research-scale eyeprocess object.
## S3 method for class 'eye_validation_job_plan' summary(object, ...)## S3 method for class 'eye_validation_job_plan' summary(object, ...)
object |
Value for 'object'. See the function description and relevant article for constraints. |
... |
Additional arguments passed to the selected engine or method. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Define synthetic measurement corruptions for stress testing
synthetic_corruption_plan( missingness = 0, pupil_dropout = 0, gaze_offset_x = 0, gaze_offset_y = 0, sampling_jitter_sd = 0, aoi_label_noise = 0, device_shift = 0, trial_drop = 0, seed = 1L )synthetic_corruption_plan( missingness = 0, pupil_dropout = 0, gaze_offset_x = 0, gaze_offset_y = 0, sampling_jitter_sd = 0, aoi_label_noise = 0, device_shift = 0, trial_drop = 0, seed = 1L )
missingness |
Generic missingness proportion. |
pupil_dropout |
Pupil dropout proportion. |
gaze_offset_x, gaze_offset_y
|
Additive gaze-coordinate offsets. |
sampling_jitter_sd |
Timestamp jitter SD in timestamp units. |
aoi_label_noise |
Proportion of AOI labels randomly reassigned. |
device_shift |
Additive shift for a declared device-sensitive numeric column. |
trial_drop |
Proportion of rows/trials removed. |
seed |
Seed. |
An object of class "eye_synthetic_corruption_plan", stored as a named list, with components "missingness", "pupil_dropout", "gaze_offset_x", "gaze_offset_y", "sampling_jitter_sd", "aoi_label_noise", "device_shift", "trial_drop", "seed", "status". It contains define synthetic measurement corruptions for stress testing and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
theory_strategy_spec(strategies = NULL, feature_columns = NULL, response = "score", participant = "participant_id", item = "item_id", condition = NULL, item_availability = NULL, engine = c("em", "stan"), multiple_starts = 10L, anchor_strength = 3, chains = 4L, parallel_chains = min(4L, chains), iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L, prototypes = NULL, feature_sd = NULL, prior = NULL)theory_strategy_spec(strategies = NULL, feature_columns = NULL, response = "score", participant = "participant_id", item = "item_id", condition = NULL, item_availability = NULL, engine = c("em", "stan"), multiple_starts = 10L, anchor_strength = 3, chains = 4L, parallel_chains = min(4L, chains), iter_warmup = 1000L, iter_sampling = 1000L, adapt_delta = 0.95, max_treedepth = 12L, prototypes = NULL, feature_sd = NULL, prior = NULL)
strategies |
Named list of strategy signatures. Each signature is a named |
feature_columns |
Process-feature columns. |
response |
Binary response column. |
participant |
Participant identifier. |
item |
Item identifier. |
condition |
Optional condition column. |
item_availability |
Optional item-by-strategy availability matrix or data frame. |
engine |
Estimation engine. |
multiple_starts |
Number of starts for the EM baseline. |
anchor_strength |
Prior/penalty strength anchoring classes to signatures. |
chains |
Stan controls. |
parallel_chains |
Stan controls. |
iter_warmup |
Stan controls. |
iter_sampling |
Stan controls. |
adapt_delta |
Stan controls. |
max_treedepth |
Stan controls. |
prototypes |
Value for 'prototypes'. See the function description and relevant article for constraints. |
feature_sd |
Value for 'feature_sd'. See the function description and relevant article for constraints. |
prior |
Value for 'prior'. See the function description and relevant article for constraints. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Timestamp semantic-fidelity audit
timestamp_fidelity_audit( source, roundtrip, source_time = "timestamp", roundtrip_time = source_time, source_unit = "seconds", roundtrip_unit = source_unit, key = NULL, tolerance = 1e-06 )timestamp_fidelity_audit( source, roundtrip, source_time = "timestamp", roundtrip_time = source_time, source_unit = "seconds", roundtrip_unit = source_unit, key = NULL, tolerance = 1e-06 )
source |
Original data. |
roundtrip |
Round-tripped data. |
source_time |
Source timestamp column. |
roundtrip_time |
Round-trip timestamp column. |
source_unit, roundtrip_unit
|
One of seconds, milliseconds, microseconds, or nanoseconds. |
key |
Optional alignment key. |
tolerance |
Seconds-scale tolerance after normalization. |
An 'eye_timestamp_fidelity' object.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
transition_residual_diagnostics(object, type = c("pearson", "deviance", "randomized"))transition_residual_diagnostics(object, type = c("pearson", "deviance", "randomized"))
object |
Dynamic IRTree fit. |
type |
Pearson, deviance, or randomized quantile residuals. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Update a partial person score with one new response
update_person_score( model, current_pattern, item_position, response, method = c("MAP", "EAP"), ... )update_person_score( model, current_pattern, item_position, response, method = c("MAP", "EAP"), ... )
model |
Calibrated model. |
current_pattern |
Existing full-length partial pattern. |
item_position |
Item receiving the new response. |
response |
New response. |
method |
Scoring method. |
... |
Additional arguments passed to the underlying method or helper. |
A named list with components "pattern", "score", containing update a partial person score with one new response and associated metadata or diagnostics.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
upgrade_eye_dataset(x, target_version = "2.0.0", copy = TRUE)upgrade_eye_dataset(x, target_version = "2.0.0", copy = TRUE)
x |
Dataset. |
target_version |
Target schema version. |
copy |
Whether to copy before migration. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
upgrade_eyeprocess_model(x, target_version = "1.0.0")upgrade_eyeprocess_model(x, target_version = "1.0.0")
x |
Model object. |
target_version |
Target model-contract version. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Compare a validation result with a frozen reference
validate_against_reference(x, reference, tolerance = 1e-06)validate_against_reference(x, reference, tolerance = 1e-06)
x |
Validation result. |
reference |
Frozen reference object or RDS path. |
tolerance |
Numeric tolerance for matched summary values. |
An object of class "eye_validation_reference_comparison", stored as a named list, with components "table", "tolerance", "pass", "reference_hash", "current_hash". It contains a validation result with a frozen reference and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validate_benchmark_study(study = eyeprocess_benchmark_study(), verify_hashes = TRUE)validate_benchmark_study(study = eyeprocess_benchmark_study(), verify_hashes = TRUE)
study |
Benchmark object or path. |
verify_hashes |
Whether to verify MD5 hashes. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Implements a lightweight contract for BIDS 1.11.1 eye-tracking physiology data. It does not replace the BIDS Validator.
validate_bids_eye_semantics(data, metadata, events_metadata = NULL)validate_bids_eye_semantics(data, metadata, events_metadata = NULL)
data |
Eye-tracking physiology table. |
metadata |
Parsed JSON sidecar as a named list. |
events_metadata |
Optional event-sidecar metadata containing 'StimulusPresentation' information for gaze-on-screen recordings. |
An 'eye_bids_semantic_audit' object.
Validate a research decision manifest
validate_decision_manifest( x, required_domains = c("sampling", "validity", "fixation", "pupil", "aoi", "model", "sensitivity", "exclusions"), require_nonempty = FALSE )validate_decision_manifest( x, required_domains = c("sampling", "validity", "fixation", "pupil", "aoi", "model", "sensitivity", "exclusions"), require_nonempty = FALSE )
x |
Manifest. |
required_domains |
Domains that must exist. |
require_nonempty |
If TRUE, required domains must contain at least one decision. |
A logical value or vector indicating a research decision manifest.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validate_engine_adapter(result, require_fit = FALSE)validate_engine_adapter(result, require_fit = FALSE)
result |
Adapter result. |
require_fit |
Require fitted status. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Validate a governed eyeprocess pipeline
validate_eye_pipeline(x)validate_eye_pipeline(x)
x |
Pipeline. |
A logical value or vector indicating a governed eyeprocess pipeline.
Validate a provenance graph
validate_eye_prov_graph(x)validate_eye_prov_graph(x)
x |
Provenance graph. |
A logical value or vector indicating a provenance graph.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validate_eye_storage_metadata(storage, verify_hashes = TRUE)validate_eye_storage_metadata(storage, verify_hashes = TRUE)
storage |
Storage object or path. |
verify_hashes |
Whether to recompute all fingerprints. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Validate that an external IRT fit used the requested engine
validate_eyeprocess_external_irt_fit(x, engine = NULL)validate_eyeprocess_external_irt_fit(x, engine = NULL)
x |
Object to validate, summarize, verify, or otherwise process. |
engine |
Requested estimation or analysis engine. |
A logical value or vector indicating that an external IRT fit used the requested engine.
Validate an adaptive IRT item bank
validate_eyeprocess_irt_item_bank(x)validate_eyeprocess_irt_item_bank(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating an adaptive IRT item bank.
Validate an eyeprocess IRT model specification
validate_eyeprocess_irt_model_spec(x)validate_eyeprocess_irt_model_spec(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating an eyeprocess IRT model specification.
Validate a joint process IRT specification
validate_eyeprocess_joint_process_irt_spec(x)validate_eyeprocess_joint_process_irt_spec(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating a joint process IRT specification.
Validate a validation-evidence plan
validate_eyeprocess_validation_plan(x)validate_eyeprocess_validation_plan(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating a validation-evidence plan.
Validate feature availability against an analysis cutoff
validate_feature_availability(provenance, cutoff)validate_feature_availability(provenance, cutoff)
provenance |
Feature provenance table. |
cutoff |
Scalar cutoff or named vector by feature. |
A data frame containing feature availability against an analysis cutoff. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Validate an integrated Gazepoint workflow result
validate_gazepoint_workflow(x)validate_gazepoint_workflow(x)
x |
An 'eye_gazepoint_workflow' object. |
A data frame of workflow checks.
This function checks presence, non-empty annotations and balanced grouping. It is deliberately a structural audit rather than a full HED validator; use the official HED validation tooling when formal schema validation is needed.
validate_hed_event_semantics(events, hed_column = "HED")validate_hed_event_semantics(events, hed_column = "HED")
events |
Value supplied to 'events'; see Details for its model-specific role. |
hed_column |
Column containing HED annotations. |
A data frame containing minimal HED annotation audit for event tables. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Validate a registered multimodal IRT model
validate_irt_model(spec, validation = NULL, ...)validate_irt_model(spec, validation = NULL, ...)
spec |
IRT model or validation specification. |
validation |
Validation results or validation specification. |
... |
Additional arguments passed to the selected model, engine, or method. |
An object of class "eye_irt_evidence_grade", stored as a named list, with components "model_id", "grade", "checks", "recovery", "contract", "warning". It contains a registered multimodal IRT model and associated metadata or diagnostics needed to interpret the result.
Validate latent-space proximity against process similarity
validate_latent_space_process_similarity( object, process_matrix, entity = c("person", "item") )validate_latent_space_process_similarity( object, process_matrix, entity = c("person", "item") )
object |
A fitted eyeprocess model or audit object. |
process_matrix |
Rows correspond to persons or items in the same order as the fitted latent coordinates. |
entity |
Entity type to map or validate. |
An object of class "eye_latent_space_process_validation", stored as a named list, with components "entity", "spearman_distance_correlation", "latent_distance", "process_distance", "interpretation". It contains latent-space proximity against process similarity and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validate_model_object(object, strict = FALSE)validate_model_object(object, strict = FALSE)
object |
Model object. |
strict |
Whether warnings become errors. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Validate a multimodal IRT development object
validate_multimodal_irt(x)validate_multimodal_irt(x)
x |
Measurement, fit, simulation, or process-information object. |
An 'eye_multimodal_validation'.
For simulations, performs data/support and identifiability checks. For fitted models, additionally audits MCMC diagnostics and optionally channel-specific posterior predictive checks.
validate_multimodal_m2(x, include_ppc = TRUE, rhat_max = 1.01, ess_min = 200)validate_multimodal_m2(x, include_ppc = TRUE, rhat_max = 1.01, ess_min = 200)
x |
M2 simulation or fit. |
include_ppc |
Include posterior predictive checks for fitted M2 models. |
rhat_max |
Maximum acceptable R-hat. |
ess_min |
Minimum bulk/tail ESS threshold. |
An 'eye_multimodal_m2_validation'.
Combines structural support, sampler diagnostics, PPC, pupil-confound availability, device/missingness summaries and explicit interpretive boundaries. Validation of synthetic or computational behavior is not empirical construct validation.
validate_multimodal_m3( x, include_ppc = TRUE, rhat_max = 1.05, ess_min = 50, ebfmi_min = 0.3 )validate_multimodal_m3( x, include_ppc = TRUE, rhat_max = 1.05, ess_min = 50, ebfmi_min = 0.3 )
x |
M3 simulation or fit. |
include_ppc |
Include M3 PPC for fits. |
rhat_max, ess_min, ebfmi_min
|
Diagnostic thresholds. |
An 'eye_multimodal_m3_validation'.
Combines structural identifiability, sampler diagnostics, state uncertainty, PPC, and optionally supplied information/negative-control/sensitivity/recovery evidence into domain-specific statuses. 'PASS' means the declared checks pass; it never means that state labels have substantive psychological validity.
validate_multimodal_m4( x, information = NULL, negative_controls = NULL, sensitivity = NULL, recovery = NULL, include_ppc = TRUE )validate_multimodal_m4( x, information = NULL, negative_controls = NULL, sensitivity = NULL, recovery = NULL, include_ppc = TRUE )
x |
M4 fit. |
information |
Optional M4 information object. |
negative_controls |
Optional M4 negative-controls object. |
sensitivity |
Optional M4 sensitivity object. |
recovery |
Optional M4 recovery object. |
include_ppc |
Whether to compute PPC summaries. |
An 'eye_multimodal_m4_validation'.
Validate a process-measure registry
validate_process_measure_registry(registry)validate_process_measure_registry(registry)
registry |
Registry data frame. |
A logical value or vector indicating a process-measure registry.
Validate a process-validation design
validate_process_validation_design(x)validate_process_validation_design(x)
x |
Validation design. |
A logical value or vector indicating a process-validation design.
Validate a process-window representation
validate_process_windows(x)validate_process_windows(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing a process-window representation. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validate_strategy_manipulation(object, condition, expected_strategy, minimum_contrast = 0)validate_strategy_manipulation(object, condition, expected_strategy, minimum_contrast = 0)
object |
Fitted strategy-mixture object. |
condition |
Condition column in the original trial data. |
expected_strategy |
Named character vector mapping condition values to prespecified strategy labels. |
minimum_contrast |
Minimum mean posterior-probability contrast over alternative strategies. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Validate imported data against a vendor semantic contract
validate_vendor_semantics(data, contract, metadata = list())validate_vendor_semantics(data, contract, metadata = list())
data |
Imported/canonical table. |
contract |
'eye_vendor_schema_contract'. |
metadata |
Optional named metadata list. |
An object of class "eye_vendor_semantic_validation", stored as a named list, with components "pass", "vendor", "version", "fields", "aliases", "timestamp", "units", "contract". It contains imported data against a vendor semantic contract and associated metadata or diagnostics needed to interpret the result.
The audit records expected semantics rather than silently coercing clocks. Tobii data can carry device/system timing; Pupil Labs Neon timestamps are high-resolution UTC nanoseconds in native recordings; Gazepoint may expose native monotonic and media-relative clocks depending on export type.
validate_vendor_timestamp_semantics( data, vendor, device_time = NULL, system_time = NULL, media_time = NULL )validate_vendor_timestamp_semantics( data, vendor, device_time = NULL, system_time = NULL, media_time = NULL )
data |
Input data frame or compatible tabular object. |
vendor |
Vendor identifier. |
device_time |
Device timestamp column. |
system_time |
System timestamp column. |
media_time |
Media/stimulus timestamp column. |
An object of class "eye_vendor_timestamp_semantics", stored as a named list, with components "vendor", "pass", "clocks". It contains vendor-specific timestamp semantics and associated metadata or diagnostics needed to interpret the result.
Evaluate a table against named validation rules
validation_acceptance_matrix(summary, rules, id_cols = character())validation_acceptance_matrix(summary, rules, id_cols = character())
summary |
Validation summary table. |
rules |
Collection of validation acceptance rules. |
id_cols |
Columns identifying validation scenarios. |
A tabular R object containing a table against named validation rules; rows represent analysis units and columns contain the returned quantities.
Define a validation acceptance rule
validation_acceptance_rule( metric, direction = c("max", "min", "between", "equals"), threshold, upper = NULL, tolerance = 0 )validation_acceptance_rule( metric, direction = c("max", "min", "between", "equals"), threshold, upper = NULL, tolerance = 0 )
metric |
Metric name or metric column. |
direction |
Direction vector used to project multidimensional information. |
threshold |
Decision or diagnostic threshold. |
upper |
Optional upper threshold for interval-style acceptance rules. |
tolerance |
Numerical or decision tolerance. |
An object of class "eye_validation_acceptance_rule", stored as a named list, with components "metric", "direction", "threshold", "upper", "tolerance". It contains define a validation acceptance rule and associated metadata or diagnostics needed to interpret the result.
Create a machine-readable validation manifest
validation_bundle_manifest(x)validation_bundle_manifest(x)
x |
Object to process, inspect, compare, or plot. |
A data frame containing a machine-readable validation manifest. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_calibration_summary(x, by = c("model_family", "scenario_id"), bins = 10L)validation_calibration_summary(x, by = c("model_family", "scenario_id"), bins = 10L)
x |
Validation collection or prediction data frame. |
by |
Grouping columns. |
bins |
Number of reliability bins. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Return stable validation condition identifiers
validation_condition_id(x)validation_condition_id(x)
x |
Validation design or expanded condition table. |
A character value or vector containing return stable validation condition identifiers.
Rank validation conditions by a transparent robustness score
validation_condition_ranking( x, weights = c(rmse = 1, abs_bias = 1, coverage_error = 1, failure_rate = 1) )validation_condition_ranking( x, weights = c(rmse = 1, abs_bias = 1, coverage_error = 1, failure_rate = 1) )
x |
Validation result. |
weights |
Named weights for rmse, absolute bias, coverage error, and failure rate. |
A tabular R object containing rank validation conditions by a transparent robustness score; rows represent analysis units and columns contain the returned quantities.
Interval-coverage table
validation_coverage_table(x, nominal = 0.95, by = NULL)validation_coverage_table(x, nominal = 0.95, by = NULL)
x |
Validation result. |
nominal |
Nominal coverage used for deviation reporting. |
by |
Optional grouping variables. |
A tabular R object containing interval-coverage table; rows represent analysis units and columns contain the returned quantities.
Returns the fine-grained evidence ladder used by the 0.7 validation programme. This is deliberately orthogonal to the existing public support levels (declared / fixture-tested / empirically-validated), so existing compatibility claims remain stable.
validation_evidence_levels()validation_evidence_levels()
A data frame ordered from weakest to strongest evidence.
Create a model-by-evidence validation matrix
validation_evidence_matrix(...)validation_evidence_matrix(...)
... |
Named validation results, bundles, or arbitrary evidence objects. |
A tabular R object containing a model-by-evidence validation matrix; rows represent analysis units and columns contain the returned quantities.
Failure profile for a validation programme
validation_failure_profile(x)validation_failure_profile(x)
x |
Validation result. |
A data frame containing failure profile for a validation programme. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_failure_summary(x, by = c("model_family", "scenario_id"))validation_failure_summary(x, by = c("model_family", "scenario_id"))
x |
Validation collection or job table. |
by |
Grouping columns. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Classify common estimator failures without hiding the original message
validation_failure_taxonomy(x)validation_failure_taxonomy(x)
x |
Error/condition/message vector. |
A data frame containing classify common estimator failures without hiding the original message. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_job_plan(grid = NULL, replications = 100L, base_seed = 1L, model_family = "unspecified", plan_id = NULL, chunk_size = 1L, metadata = list())validation_job_plan(grid = NULL, replications = 100L, base_seed = 1L, model_family = "unspecified", plan_id = NULL, chunk_size = 1L, metadata = list())
grid |
Scenario grid or named list of factor levels. |
replications |
Replications per design cell. |
base_seed |
Base seed used for deterministic seed allocation. |
model_family |
Model-family label. |
plan_id |
Optional explicit plan identifier. |
chunk_size |
Number of jobs assigned to each chunk. |
metadata |
Arbitrary plan metadata. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Build a measurement-to-generalization validation ladder
validation_ladder( acquisition_qc = "not_assessed", analytical_qc = "not_assessed", construct_check = "not_assessed", within_person = "not_assessed", held_out_person = "not_assessed", claim = "descriptive" )validation_ladder( acquisition_qc = "not_assessed", analytical_qc = "not_assessed", construct_check = "not_assessed", within_person = "not_assessed", held_out_person = "not_assessed", claim = "descriptive" )
acquisition_qc, analytical_qc, construct_check, within_person, held_out_person
|
Stage statuses: 'pass', 'warning', 'fail', or 'not_assessed'. |
claim |
Claim type; use 'generalizable' for out-of-person claims. |
A structured validation-ladder result.
Monte Carlo standard errors for validation metrics
validation_mcse(results, metric = c("bias", "rmse", "coverage"))validation_mcse(results, metric = c("bias", "rmse", "coverage"))
results |
Validation or model results. |
metric |
Metric to calculate or audit. |
A data frame containing monte Carlo standard errors for validation metrics. Rows represent the analysis units and columns contain the identifiers, estimates, or diagnostics defined by the function.
Estimate Monte Carlo uncertainty for validation summaries
validation_mcse_profile(x, metric, by = character())validation_mcse_profile(x, metric, by = character())
x |
Object to validate, summarize, verify, or otherwise process. |
metric |
Metric name or metric column. |
by |
Grouping variables or aggregation level. |
A tabular R object containing monte Carlo uncertainty for validation summaries; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_recovery_summary(x, by = character())validation_recovery_summary(x, by = character())
x |
Validation collection or estimates data frame. |
by |
Additional grouping columns. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Parameter-recovery table
validation_recovery_table(x, by = NULL)validation_recovery_table(x, by = NULL)
x |
Validation result. |
by |
Optional grouping variables. |
A tabular R object containing parameter-recovery table; rows represent analysis units and columns contain the returned quantities.
Compute a replication budget from a target MCSE
validation_replication_budget( pilot_sd, target_mcse, minimum = 20L, maximum = 10000L )validation_replication_budget( pilot_sd, target_mcse, minimum = 20L, maximum = 10000L )
pilot_sd |
Pilot estimate of the metric standard deviation. |
target_mcse |
Target Monte Carlo standard error. |
minimum |
Minimum permitted replication count. |
maximum |
Maximum permitted replication count. |
A numeric value or vector containing a replication budget from a target MCSE.
Render a conservative validation report
validation_report(x, include_session = TRUE)validation_report(x, include_session = TRUE)
x |
Validation bundle. |
include_session |
Include abbreviated session provenance. |
Character vector of report lines.
Overall validation robustness score
validation_robustness_score(x)validation_robustness_score(x)
x |
Validation result. |
A single numeric robustness score: the mean finite condition-level robustness score, or 'NA_real_' when no finite score is available.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_runtime_summary(x, by = c("model_family", "scenario_id"))validation_runtime_summary(x, by = c("model_family", "scenario_id"))
x |
Validation collection or job table. |
by |
Grouping columns. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_sbc_summary(x, by = c("model_family", "scenario_id"), bins = 10L)validation_sbc_summary(x, by = c("model_family", "scenario_id"), bins = 10L)
x |
Validation collection or posterior draws data frame. |
by |
Grouping columns. |
bins |
Rank-histogram bins. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Create a scenario manifest for frozen validation work
validation_scenario_manifest( plan, source_commit = NA_character_, generated_at = Sys.time() )validation_scenario_manifest( plan, source_commit = NA_character_, generated_at = Sys.time() )
plan |
Validation or stress-evidence plan object. |
source_commit |
Source-control commit associated with the evidence. |
generated_at |
Generation timestamp stored in the manifest. |
An object of class "eye_validation_scenario_manifest", stored as a named list, with components "label", "plan_hash", "scenarios", "source_commit", "generated_at", "scientific_scope". It contains a scenario manifest for frozen validation work and associated metadata or diagnostics needed to interpret the result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_seed(design, replication, base_seed = 1L, stream = 1L)validation_seed(design, replication, base_seed = 1L, stream = 1L)
design |
A list or one-row data frame describing a design cell. |
replication |
Positive replication number. |
base_seed |
Base integer seed. |
stream |
Optional independent stream number. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Monte Carlo standard-error diagnostics for validation summaries
validation_summary_mcse(x, by = NULL)validation_summary_mcse(x, by = NULL)
x |
Validation result. |
by |
Optional grouping variables. |
A tabular R object containing monte Carlo standard-error diagnostics for validation summaries; rows represent analysis units and columns contain the returned quantities.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
validation_thresholds(required_replications = 100L, max_failure_rate = 0.05, max_absolute_bias = 0.10, max_rmse = Inf, min_coverage = 0.90, max_coverage = 0.99, max_rhat = 1.01, min_ess_bulk = 400, max_divergence_rate = 0.01, require_sbc = TRUE, require_empirical_reproduction = TRUE)validation_thresholds(required_replications = 100L, max_failure_rate = 0.05, max_absolute_bias = 0.10, max_rmse = Inf, min_coverage = 0.90, max_coverage = 0.99, max_rhat = 1.01, min_ess_bulk = 400, max_divergence_rate = 0.01, require_sbc = TRUE, require_empirical_reproduction = TRUE)
required_replications |
Required completed replications per scenario. |
max_failure_rate |
Maximum tolerated job failure rate. |
max_absolute_bias |
Maximum absolute mean bias. |
max_rmse |
Maximum RMSE. |
min_coverage |
Minimum interval coverage. |
max_coverage |
Maximum interval coverage. |
max_rhat |
Maximum acceptable R-hat. |
min_ess_bulk |
Minimum bulk effective sample size. |
max_divergence_rate |
Maximum divergence rate. |
require_sbc |
Require passing SBC uniformity evidence. |
require_empirical_reproduction |
Require empirical-reproduction evidence. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Declare a vendor semantic schema contract
vendor_schema_contract( vendor, version = NA_character_, required_fields = character(), optional_fields = character(), aliases = list(), timestamp = list(), coordinate = list(), units = list(), eye_streams = character(), event_fields = character() )vendor_schema_contract( vendor, version = NA_character_, required_fields = character(), optional_fields = character(), aliases = list(), timestamp = list(), coordinate = list(), units = list(), eye_streams = character(), event_fields = character() )
vendor |
Vendor/ecosystem label. |
version |
Optional format/software version. |
required_fields |
Fields that must survive import. |
optional_fields |
Fields that may be present. |
aliases |
Named list mapping canonical fields to accepted vendor names. |
timestamp |
Named list describing device/system/media time columns. |
coordinate |
Named list describing x/y columns and coordinate semantics. |
units |
Named list of expected units for canonical fields. |
eye_streams |
Expected eye streams ('left', 'right', 'cyclopean', etc.). |
event_fields |
Event/annotation fields expected to survive. |
An object of class "eye_vendor_schema_contract", stored as a named list, with components "vendor", "version", "required_fields", "optional_fields", "aliases", "timestamp", "coordinate", "units", "eye_streams", "event_fields", "contract_version". It contains declare a vendor semantic schema contract and associated metadata or diagnostics needed to interpret the result.
Specify multi-vendor empirical validation requirements
vendor_validation_spec( required_vendors = c("gazepoint", "tobii", "pupillabs", "eyelink", "smi"), min_cases_per_vendor = 2L, min_pass_rate = 0.95, require_versions = TRUE, require_devices = TRUE, require_independent_sources = TRUE, require_licence_reviewed = TRUE )vendor_validation_spec( required_vendors = c("gazepoint", "tobii", "pupillabs", "eyelink", "smi"), min_cases_per_vendor = 2L, min_pass_rate = 0.95, require_versions = TRUE, require_devices = TRUE, require_independent_sources = TRUE, require_licence_reviewed = TRUE )
required_vendors |
Vendors that must be represented. |
min_cases_per_vendor |
Minimum independent cases per vendor. |
min_pass_rate |
Minimum acceptable pass rate per vendor. |
require_versions |
Require non-missing software versions. |
require_devices |
Require non-missing device models. |
require_independent_sources |
Require each case to be marked as an independently obtained source rather than a duplicated fixture. |
require_licence_reviewed |
Require each corpus case to have a completed data/code licence and redistribution review. |
An 'eye_vendor_validation_spec'.
Verify that a locked manifest has not changed
verify_decision_manifest_lock(x)verify_decision_manifest_lock(x)
x |
Manifest lock. |
A logical value or vector indicating verify that a locked manifest has not changed.
Verify a frozen validation atlas
verify_eyeprocess_validation_atlas(x)verify_eyeprocess_validation_atlas(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating verify a frozen validation atlas.
Verify the integrity hash of a frozen evidence bundle
verify_eyeprocess_validation_evidence(x)verify_eyeprocess_validation_evidence(x)
x |
Object to validate, summarize, verify, or otherwise process. |
A logical value or vector indicating verify the integrity hash of a frozen evidence bundle.
Verify an outcome-blind snapshot has not changed
verify_outcome_blind_snapshot(x)verify_outcome_blind_snapshot(x)
x |
Snapshot. |
A logical value or vector indicating verify an outcome-blind snapshot has not changed.
Verify an internally stored fingerprint hash
verify_reproducibility_fingerprint(x)verify_reproducibility_fingerprint(x)
x |
Fingerprint. |
A logical value or vector indicating verify an internally stored fingerprint hash.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
verify_reproducibility_manifest(manifest)verify_reproducibility_manifest(manifest)
manifest |
Manifest. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Build an item-to-visual-context registry
visual_context_registry( item_metadata, item = "item_id", context = NULL, context_candidates = c("visual_anchor_id", "stimulus_id", "stimulus_page", "page_id", "layout_id", "screen_id", "diagram_id"), min_items_per_context = 3L )visual_context_registry( item_metadata, item = "item_id", context = NULL, context_candidates = c("visual_anchor_id", "stimulus_id", "stimulus_page", "page_id", "layout_id", "screen_id", "diagram_id"), min_items_per_context = 3L )
item_metadata |
Item-level metadata. |
item |
Item identifier column. |
context |
Optional explicit context column. |
context_candidates |
Candidate metadata columns searched in order. |
min_items_per_context |
Minimum items required for a shared context. |
An object of class "eye_visual_context_registry", stored as a named list, with components "mapping", "source_item_column", "source_context_column", "min_items_per_context", "caveat". It contains an item-to-visual-context registry and associated metadata or diagnostics needed to interpret the result.
Write an advanced-model evidence report
write_advanced_model_evidence_report(x, path)write_advanced_model_evidence_report(x, path)
x |
An 'eye_advanced_evidence_audit' object. |
path |
Markdown output file. |
Normalized report path.
Write API lifecycle registry to CSV
write_api_lifecycle_registry(registry, path)write_api_lifecycle_registry(registry, path)
registry |
Lifecycle registry. |
path |
Output path. |
An R object containing aPI lifecycle registry to CSV. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_benchmark_data_dictionary(study = eyeprocess_benchmark_study(), path = "benchmark-data-dictionary.md")write_benchmark_data_dictionary(study = eyeprocess_benchmark_study(), path = "benchmark-data-dictionary.md")
study |
Benchmark object or path. |
path |
Output Markdown file. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Write a decision manifest
write_decision_manifest(x, path, format = c("rds", "dput", "json"))write_decision_manifest(x, path, format = c("rds", "dput", "json"))
x |
Manifest. |
path |
Output path. |
format |
'rds', 'dput', or 'json'. |
An R object containing a decision manifest. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Write a conservative pipeline report
write_eye_pipeline_report(x, path)write_eye_pipeline_report(x, path)
x |
Pipeline or run. |
path |
Output text/markdown path. |
A character string or vector giving the path or identifier for a conservative pipeline report.
Write an eye dataset to RDS or Arrow/Parquet storage
write_eye_storage( x, path, format = c("rds", "parquet", "arrow_dataset"), tables = canonical_table_names(), partitioning = NULL, compression = "zstd", overwrite = FALSE, retain_metadata = TRUE )write_eye_storage( x, path, format = c("rds", "parquet", "arrow_dataset"), tables = canonical_table_names(), partitioning = NULL, compression = "zstd", overwrite = FALSE, retain_metadata = TRUE )
x |
An 'eye_dataset'. |
path |
Output path. |
format |
Storage format. |
tables |
Canonical tables to write. |
partitioning |
Optional partition columns for Arrow datasets. |
compression |
Parquet compression codec. For writes, the default '"zstd"' is preferred; when the argument is omitted and that codec is unavailable, storage falls back to '"snappy"' and then '"uncompressed"'. An explicitly requested unavailable codec errors. |
overwrite |
Whether to replace an existing target. |
retain_metadata |
Whether to retain raw/vendor metadata in a sidecar RDS. |
An 'eye_storage' handle.
Write an explicit '_targets.R' template from a governed pipeline
write_eye_targets_template(x, path = "_targets.R")write_eye_targets_template(x, path = "_targets.R")
x |
Pipeline. |
path |
Output path. |
An R object containing an explicit '_targets.R' template from a governed pipeline. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Write a frozen evidence bundle
write_eyeprocess_validation_evidence(x, path)write_eyeprocess_validation_evidence(x, path)
x |
Object to validate, summarize, verify, or otherwise process. |
path |
File path for reading or writing. |
A character string or vector giving the path or identifier for a frozen evidence bundle.
Write a compact Markdown validation report
write_eyeprocess_validation_report( atlas, path, title = "eyeprocess validation evidence report" )write_eyeprocess_validation_report( atlas, path, title = "eyeprocess validation evidence report" )
atlas |
Validation-evidence atlas object. |
path |
File path for reading or writing. |
title |
Report title. |
A character string or vector giving the path or identifier for a compact Markdown validation report.
Write a reproducible Gazepoint workflow report
write_gazepoint_workflow_report( workflow, path = file.path(workflow$output_dir, "gazepoint-workflow-report.md"), render_html = workflow$spec$create_html_report )write_gazepoint_workflow_report( workflow, path = file.path(workflow$output_dir, "gazepoint-workflow-report.md"), render_html = workflow$spec$create_html_report )
workflow |
An 'eye_gazepoint_workflow' result. |
path |
Markdown report destination. |
render_html |
Render an HTML copy when possible. |
The normalized report path.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_model_promotion_report(x, path)write_model_promotion_report(x, path)
x |
Model-promotion audit. |
path |
Markdown output file. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_partitioned_eye_storage(x, path, spec = partition_eye_storage(format = if (requireNamespace("arrow", quietly = TRUE)) "parquet" else "rds"), overwrite = FALSE, tables = NULL)write_partitioned_eye_storage(x, path, spec = partition_eye_storage(format = if (requireNamespace("arrow", quietly = TRUE)) "parquet" else "rds"), overwrite = FALSE, tables = NULL)
x |
Eye dataset or named list of data frames. |
path |
Destination directory. |
spec |
Partition specification. |
overwrite |
Whether to replace an existing store. |
tables |
Tables to write. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Return Graphviz DOT for a provenance graph
write_prov_dot(x)write_prov_dot(x)
x |
Provenance graph. |
A character value or vector containing return Graphviz DOT for a provenance graph.
Write a reporting-guideline audit report
write_reporting_guideline_report(x, path)write_reporting_guideline_report(x, path)
x |
An 'eye_reporting_audit' object. |
path |
Markdown output file. |
Normalized report path.
Write a reproducibility fingerprint
write_reproducibility_fingerprint(x, path, format = c("rds", "dput", "json"))write_reproducibility_fingerprint(x, path, format = c("rds", "dput", "json"))
x |
Fingerprint. |
path |
Output path. |
format |
'rds', 'dput', or 'json'. |
A character string or vector giving the path or identifier for a reproducibility fingerprint.
Write a human-readable software-paper evidence report
write_software_paper_evidence(x, path)write_software_paper_evidence(x, path)
x |
Evidence bundle. |
path |
Markdown output path. |
An R object containing a human-readable software-paper evidence report. The concrete class and structure follow the selected method, engine, or input object and are preserved as documented by that workflow.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_software_paper_reproduction(directory, study = eyeprocess_benchmark_study(), overwrite = FALSE)write_software_paper_reproduction(directory, study = eyeprocess_benchmark_study(), overwrite = FALSE)
directory |
Output directory. |
study |
Benchmark study. |
overwrite |
Whether to replace existing files. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Write a methodological software-paper scaffold
write_software_paper_scaffold( path, title = "eyeprocess: Reproducible Psychometric Process Modelling in R", author = "Stefanos Balaskas" )write_software_paper_scaffold( path, title = "eyeprocess: Reproducible Psychometric Process Modelling in R", author = "Stefanos Balaskas" )
path |
Output R Markdown file. |
title |
Paper title. |
author |
Author string. |
The normalized path.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_validation_job_manifest(plan, path, overwrite = FALSE)write_validation_job_manifest(plan, path, overwrite = FALSE)
plan |
An 'eye_validation_job_plan'. |
path |
Manifest directory. |
overwrite |
Replace an existing manifest directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_validation_release_report(x, path, completion = NULL, promotion = NULL, title = "eyeprocess validation release report", include_session = TRUE)write_validation_release_report(x, path, completion = NULL, promotion = NULL, title = "eyeprocess validation release report", include_session = TRUE)
x |
Validation collection. |
path |
Markdown output file. |
completion |
Optional completion audit. |
promotion |
Optional model-promotion audit. |
title |
Report title. |
include_session |
Include session information. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Write a validation report to disk
write_validation_report(x, path, ...)write_validation_report(x, path, ...)
x |
Validation bundle. |
path |
Output text-file path. |
... |
Passed to 'validation_report()'. |
A character string or vector giving the path or identifier for a validation report to disk.
Write a validation scenario manifest
write_validation_scenario_manifest(x, path)write_validation_scenario_manifest(x, path)
x |
Object to validate, summarize, verify, or otherwise process. |
path |
File path for reading or writing. |
A character string or vector giving the path or identifier for a validation scenario manifest.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_vendor_case_report(corpus_path, case_id, path, validation = NULL, roundtrip = NULL)write_vendor_case_report(corpus_path, case_id, path, validation = NULL, roundtrip = NULL)
corpus_path |
Corpus directory. |
case_id |
Case identifier. |
path |
Markdown output path. |
validation |
Optional validation result. |
roundtrip |
Optional round-trip audit. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Part of the research-scale validation, advanced-model, interoperability, storage, adapter, or reproducibility programme. Experimental model functions remain subject to declared evidence gates.
write_vendor_registry(x, corpus_path)write_vendor_registry(x, corpus_path)
x |
Registry data frame. |
corpus_path |
Corpus directory. |
The documented eyeprocess object, data frame, plot, report path, or adapter result.
Write a multi-vendor validation report
write_vendor_validation_report(x, path)write_vendor_validation_report(x, path)
x |
An 'eye_vendor_validation' object. |
path |
Markdown output file. |
The normalized report path.