TRexScreeningSelector: T-Rex Screening Selector (T-Rex+Screening)

TRexScreeningSelectorR Documentation

T-Rex Screening Selector (T-Rex+Screening)

Description

Fast FDR-controlled variable screening in one pass of the random experiments, until one dummy is included (T = 1), with as many dummies as variables (L = p).

Fields inherited from TRexSelector

The screening has no target FDR and no calibration, so the fields of TRexSelector hold fixed values after a run: T_stop is 1, L is 1, num_dummies is p, v_thresh is 0.5 and tFDR is 0. r_mat is a 1 x 1 matrix with the number of selected variables, phi_prime equals the relative occurrences, and fdp_hat_mat is a 1 x 1 matrix with 0, because no FDP is estimated per voting level. The FDR estimate of the run is estimated_FDR.

A response without signal

The screening stops at the first dummy, so about one null variable is expected to enter before it. On a response of pure noise it selected at least one variable in a quarter to a half of the runs in simulations, with fewer than one selected variable on average. A small selection is therefore no evidence of a signal. The estimate, one over the number of selected variables, rests on that expected number of false discoveries and is no bound on the FDP of a run: with two or more variables selected on noise it is 0.5 or less, while every one of them is false. TRexBiobankScreeningSelector accepts a screening result only when the estimate lies in its window and runs the T-Rex selector otherwise, which selected nothing on noise in about 99 % of the runs.

L-loop strategies

The screening takes the strategies "STANDARD" (default), "SEEDED", "PERMUTATION" and "PERMUTATION_SEEDED" of trex_control(). "SEEDED" gives the results of "STANDARD" and keeps one dummy matrix in memory at a time. "PERMUTATION_SEEDED" gives the results of "PERMUTATION" without storing the base dummy matrix. "HCONCAT" and "SKIPL" are refused, because they describe the growth of the dummy matrix over the L-loop, and the screening is one pass at L = p.

Super class

TRexSelector -> TRexScreeningSelector

Active bindings

estimated_FDR

The estimated FDR, one over the number of selected variables. An empty selection reports 1. It rests on the expected number of false discoveries, about one, and is no bound on the FDP of a run (see the section "A response without signal").

used_bootstrap

Whether the confidence-based screening (bootstrap) made the selection.

confidence_level

Confidence level that the grid search of the bootstrap picked. 0 without the bootstrap.

estimated_correlation

Correlation coefficient of the dependency-aware variants: the supplied value, or the estimate from the data. 0 for "TREX".

beta_mat

Coefficients of the variables, averaged over the K experiments (one per variable, a p x 1 matrix).

dummy_betas

Non-zero coefficients of the dummies, collected over the K experiments. They are the sample of the bootstrap.

Methods

Public methods

Inherited methods

TRexScreeningSelector$new()

Create a new TRexScreeningSelector object.

Usage
TRexScreeningSelector$new(
  X,
  y,
  seed = -1,
  verbose = TRUE,
  screen_control = trex_screen_control(),
  control = trex_control()
)
Arguments
X

Feature matrix in memory, or an mmap_matrix, the handle of a memory-mapped file. Both are bound without a copy. The selector standardizes a memory-mapped design in place and restores it, which needs "readwrite" mode. Under scaling_mode = "NONE" it writes nothing and "readonly" mode suffices.

y

Response vector.

seed

Random seed. A negative value (default: -1) gives a run that is not reproducible.

verbose

Whether to print progress (default: TRUE).

screen_control

A screen control list from trex_screen_control() (default: trex_screen_control()).

control

A control list from trex_control() (default: trex_control()). See the section on the L-loop strategies for the values of lloop_strategy that the screening takes.


TRexScreeningSelector$select()

Run the Screening algorithm. A selector runs once: a second call is an error, see TRexSelector.

Usage
TRexScreeningSelector$select()

TRexScreeningSelector$clone()

The objects of this class are cloneable with this method.

Usage
TRexScreeningSelector$clone(deep = FALSE)
Arguments
deep

Whether to make a deep clone.

References

J. Machkour, M. Muma, and D. P. Palomar, "False Discovery Rate Control for Fast Screening of Large-Scale Genomics Biobanks," in Proc. 22nd IEEE Statist. Signal Process. Workshop (SSP), 2023, pp. 666-670.

J. Machkour, M. Muma, and D. P. Palomar, "The Terminating-Random Experiments Selector: Fast High-Dimensional Variable Selection with False Discovery Rate Control," Signal Process., vol. 231, Art. no. 109894, 2025, doi: 10.1016/j.sigpro.2025.109894.

See Also

Other T-Rex selectors and controls: Gauss_data, TRexBiobankScreeningSelector, TRexDASelector, TRexGVSSelector, TRexSPCASelector, TRexSelector, TRexTikhonovSelector, trex_biobank_control(), trex_build_group_centering_k(), trex_control(), trex_da_control(), trex_default_lambda2_grid(), trex_dummy_distribution(), trex_fdp(), trex_fdp_dense(), trex_gamma_to_k(), trex_get_num_threads(), trex_gvs_control(), trex_gvs_lambda2_power_rule(), trex_normalize_design(), trex_precision(), trex_recall(), trex_screen_control(), trex_spca_control(), trex_tikhonov_control(), trex_tikhonov_lambda2_power_rule(), trex_tpp(), trex_tpp_dense()

Examples


data(Gauss_data)
X <- Gauss_data$X
y <- Gauss_data$y
sel <- TRexScreeningSelector$new(X, y, verbose = FALSE)
sel$select()
sel$selected_indices



TRexSelector documentation built on Nov. 20, 2026, 5:09 p.m.