Computational benchmarking and synthetic stress testing

In eyeprocess: Harmonize Eye-Tracking, Pupillometry, Biometrics, and Psychometric Process Data

knitr::opts_chunk$set(collapse = TRUE, comment = "#>")

Scientific validity and computational feasibility are separate questions. eye_benchmark_design() measures runtime/scaling under declared dataset sizes, while synthetic corruption plans probe robustness to missingness, pupil dropout, calibration offsets, timestamp jitter, AOI label noise, device shifts, and trial imbalance.

plans <- list(
 synthetic_corruption_plan(missingness=.05),
 synthetic_corruption_plan(missingness=.20, sampling_jitter_sd=2),
 synthetic_corruption_plan(pupil_dropout=.30, gaze_offset_x=.02)
)
st <- stress_test_process_pipeline(data, plans, analysis_fun)
stress_test_summary(st)
plot(st, severity="missingness", metric="effect")

Stress tests describe sensitivity to the perturbations actually supplied. They do not replace validation on independent empirical data.

eyeprocess documentation built on Sept. 28, 2026, 5:08 p.m.