Configure group examination

examination_control() validates the bounded execution, residual-geometry, review, staging, and numerical settings used by examine_group(). Keeping these settings in a typed object makes defaults visible and reproducible.

Usage

examination_control(
  block_size = 1024L,
  geometry = list(rank = 32L, oversample = 8L, cap = 8, balance_contrasts = TRUE,
    stability_replicates = 2L),
  review = list(surprise = list(energy_threshold = 2.5, tail_threshold = 0.01,
    residual_threshold = 3), influence = list(energy_threshold = 1, max_abs_threshold =
    2), quality = list(coverage_fraction = list(direction = "low", threshold = 0.8)),
    min_stability = 0.7),
  exact_refit_n = 5L,
  retain_n = 5L,
  staging = list(tempdir = NULL),
  tolerance = list(rank = sqrt(.Machine$double.eps), leverage = 1e-08, degeneracy =
    1e-12)
)

Arguments

block_size

Positive number of samples read per scan block.

geometry

Named list controlling the deterministic low-rank residual representation. Supported entries are rank, oversample, cap, balance_contrasts, and stability_replicates. With balance_contrasts = TRUE each contrast's residuals are scaled by its own eligible feature count so contrasts contribute comparably to the geometry. stability_replicates = 0 disables split-feature stability assessment; review criteria are then not gated on stability.

review

Named list controlling review status. Supported entries are surprise, influence, quality, and min_stability.

exact_refit_n

Maximum number of retained random-effects subjects, taken in decreasing review_priority, that receive an exact heterogeneity refit. Other retained subjects keep screening-mode maps.

retain_n

Number of highest-priority subjects retained for localization when no explicit retain set is supplied.

staging

Named list with optional tempdir. Staging stores are always removed on success or failure.

tolerance

Named list of numerical tolerances: rank, leverage, and degeneracy.

Value

An object of class gds_examination_control.