cca_rrr: Canonical Correlation Analysis via Reduced Rank Regression...

In ccar3: Canonical Correlation Analysis via Reduced Rank Regression

Canonical Correlation Analysis via Reduced Rank Regression (RRR)

Description

Estimates canonical directions using various RRR solvers and penalties.

Usage

cca_rrr(
  X,
  Y,
  Sx = NULL,
  Sy = NULL,
  lambda = 0,
  r,
  highdim = TRUE,
  solver = "ADMM",
  LW_Sy = TRUE,
  standardize = TRUE,
  rho = 1,
  niter = 10000,
  thresh = 1e-04,
  thresh_0 = 1e-06,
  verbose = FALSE
)

Arguments

X	Matrix of predictors.
Y	Matrix of responses.
Sx	Optional X covariance matrix.
Sy	Optional Y covariance matrix.
lambda	Regularization parameter.
r	Rank of the solution.
highdim	Boolean for high-dimensional regime.
solver	Solver type: "rrr", "CVX", or "ADMM".
LW_Sy	Whether to use Ledoit-Wolf shrinkage for Sy.
standardize	Logical; should X and Y be scaled.
rho	ADMM parameter.
niter	Maximum number of iterations for ADMM.
thresh	Convergence threshold.
thresh_0	For the ADMM solver: Set entries whose absolute value is below this to 0 (default 1e-6).
verbose	Logical for verbose output.

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.