trex_spca_control: T-Rex+SPCA Control Parameters

View source: R/trex_spca_selector.R

trex_spca_controlR Documentation

T-Rex+SPCA Control Parameters

Description

Build the control list of the T-Rex sparse PCA selector (T-Rex+SPCA), TRexSPCASelector. The list holds the parameters below. Every other parameter of the T-Rex+GVS run of each component (the clustering, the cross-validation grid, the L-loop strategy, the scaling mode, the threads) runs at its default, see trex_gvs_control() and trex_control().

Usage

trex_spca_control(
  mode = "ActiveSet",
  lambda2_ridge_loadings = 1e-06,
  gvs_type = "EN",
  lambda_2 = -1,
  lambda2_method = "CV_1SE_CCD",
  en_solver = "TENET_AUG",
  tenet_aug_use_lars = FALSE,
  K = 20L,
  max_dummy_multiplier = 10L
)

Arguments

mode

Assembly of the sparse loading vectors. "ActiveSet" (default) is the ridge regression of the component on the selected variables, normalized to unit length (the T-Rex PCA of the paper). "Thresholded" keeps the largest ordinary PCA loadings in absolute value, as many as variables were selected, normalized to unit length (the T-Rex Thresholded PCA of the paper).

lambda2_ridge_loadings

Ridge penalty of the loading assembly in "ActiveSet" mode (default: 1e-6). Not used in "Thresholded" mode.

gvs_type

Penalty of the T-Rex+GVS runs. Only "EN" (default, elastic net) is accepted: the method is defined on the elastic net, and "IEN" is refused.

lambda_2

Ridge part of the elastic net of the T-Rex+GVS runs. A negative value (default: -1) lets the rule lambda2_method choose it, 0 is the T-LASSO without the ridge part, and a positive value is used as it is. A supplied value refers to the scale of the standardized columns.

lambda2_method

Rule that chooses lambda_2 when lambda_2 < 0: "CV_1SE_CCD" (default), "CV_MIN_CCD", "CV_1SE_SVD" or "CV_MIN_SVD". The rules of the informed elastic net are refused.

en_solver

Elastic net solver of the T-Rex+GVS runs. "TENET_AUG" (default) is the elastic net as a LASSO on row-augmented data, the form of the paper. "TENET" is the elastic net on the Gram matrix. Both solve the same elastic net problem.

tenet_aug_use_lars

Logical. For en_solver = "TENET_AUG": run the augmented system with a pure-LARS inner path that never drops variables (TRUE) instead of the LASSO inner path that can (FALSE, default). Ignored for "TENET".

K

Number of random experiments of each T-Rex+GVS run (default: 20).

max_dummy_multiplier

Cap of the dummy multiplier of each T-Rex+GVS run: the L-loop adds dummies up to this many times the number of variables (default: 10).

Value

A named list for passing to TRexSPCASelector$new().

See Also

Other T-Rex selectors and controls: Gauss_data, TRexBiobankScreeningSelector, TRexDASelector, TRexGVSSelector, TRexSPCASelector, TRexScreeningSelector, 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_tikhonov_control(), trex_tikhonov_lambda2_power_rule(), trex_tpp(), trex_tpp_dense()

Examples

trex_spca_control()
trex_spca_control(mode = "Thresholded", K = 10)
trex_spca_control(lambda_2 = 0.01)


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