cca_group_rrr_cv: Group-Sparse Canonical Correlation via Reduced-Rank...

In ccar3: Canonical Correlation Analysis via Reduced Rank Regression

Group-Sparse Canonical Correlation via Reduced-Rank Regression with CV

Description

Performs group-sparse reduced-rank regression for CCA using either ADMM or CVXR solvers.

Usage

cca_group_rrr_cv(
  X,
  Y,
  groups,
  r = 2,
  lambdas = 10^seq(-3, 1.5, length.out = 10),
  kfolds = 5,
  parallelize = FALSE,
  standardize = FALSE,
  LW_Sy = TRUE,
  solver = "ADMM",
  rho = 1,
  thresh_0 = 1e-06,
  niter = 10000,
  thresh = 1e-04,
  verbose = FALSE,
  nb_cores = NULL
)

Arguments

X	Predictor matrix (n x p)
Y	Response matrix (n x q)
groups	List of index vectors defining groups of predictors
r	Target rank
lambdas	Grid of regularization parameters to try out
kfolds	Nb of folds for the CV procedure
parallelize	Whether to use parallel processing (default is FALSE)
standardize	Whether to scale variables
LW_Sy	Whether to apply Ledoit-Wolf shrinkage to Sy (default TRUE)
solver	Either "ADMM" or "CVXR"
rho	ADMM parameter
thresh_0	tolerance for declaring entries non-zero
niter	Maximum number of ADMM iterations
thresh	Convergence threshold for ADMM
verbose	Print diagnostics
nb_cores	Number of cores to use for parallelization (default is all available cores minus 1)

Value

A list with elements:

U

Canonical direction matrix for X (p x r)

V

Canonical direction matrix for Y (q x r)

lambda

Optimal regularisation parameter lambda chosen by CV

rmse

Mean squared error of prediction (as computed in the CV)

cor

Canonical covariances

ccar3 documentation built on Sept. 16, 2025, 9:11 a.m.