Skip to contents

Run ICA with cross-validation and random initialisation

rs_ica_iters_cv.Rd

[Experimental] This function will split the data into no_folds and apply ICA with no_random_init random initialisations over each fold.

Usage

rs_ica_iters_cv(
  x,
  no_comp,
  no_folds,
  no_random_init,
  ica_type,
  random_seed,
  ica_params
)

Arguments

x

Numeric matrix. The processed data (no whitening function has been applied yet.)

no_comp

Integer. Number of components to test for.

no_folds

Integer. Number of folds to use for the cross-validation.

no_random_init

Integer. Number of random initialisations per fold.

ica_type

String. One of c("logcosh", "exp").

random_seed

Integer. For reproducibility.

ica_params

A list containing:

  • maxit - Integer. Maximum number of iterations for ICA.

  • alpha - Float. The alpha parameter for the logcosh version of ICA. Should be between 1 to 2.

  • max_tol - Maximum tolerance of the algorithm

  • verbose - Verbosity of the function, i.e., shall individual iters be shown.

If the list is empty or the expected elements are not found, default values are used.

Value

A list containing:

  • s_combined - The combined matrices for S. Dimensions are nrows = features; and ncols = no_comp * no_random_init * no_folds.

  • converged - Boolean vector indicating if the respective run reached convergence. Length = no_random_init * no_folds