ecca: Sparse Canonical Correlation via Reduced-Rank Regression when...

In ccar3: Canonical Correlation Analysis via Reduced Rank Regression

Sparse Canonical Correlation via Reduced-Rank Regression when both X and Y are high-dimensional.

Description

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

Usage

ecca(
  X,
  Y,
  lambda = 0,
  groups = NULL,
  Sx = NULL,
  Sy = NULL,
  Sxy = NULL,
  r = 2,
  standardize = FALSE,
  rho = 1,
  B0 = NULL,
  eps = 1e-04,
  maxiter = 500,
  verbose = TRUE
)

Arguments

X	Predictor matrix (n x p)
Y	Response matrix (n x q)
lambda	Regularization parameter
groups	List of index vectors defining groups of predictors
Sx	precomputed covariance matrix for X (optional)
Sy	precomputed covariance matrix for Y (optional)
Sxy	precomputed covariance matrix between X and Y (optional)
r	Target rank
standardize	Whether to scale variables
rho	ADMM parameter
B0	Initial value for the coefficient matrix (optional)
eps	Convergence threshold for ADMM
maxiter	Maximum number of ADMM iterations
verbose	Print diagnostics

Value

A list with elements:

U

Canonical direction matrix for X (p x r)

V

Canonical direction matrix for Y (q x r)

cor

Canonical covariances

loss

The prediction error 1/n * \| XU - YV\|^2

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