Compute a ridge regression solution path — ridge_path • amatrix

Fits ridge regression for every penalty value in lambdas via a single thin SVD of X, returning coefficients for all penalties at once.

Usage

ridge_path(X, Y, lambdas, k = NULL, ...)

Arguments

X: Numeric matrix or adgeMatrix of predictors, shape [n, p].
Y: Numeric matrix, vector, or adgeMatrix of responses, shape [n, q].
lambdas: Positive numeric vector of ridge penalty values. Must satisfy all(lambdas > 0).
k: Integer or NULL. Number of singular values to retain in the truncated SVD. When NULL, defaults to min(nrow(X), ncol(X)).
...: Additional arguments forwarded to svd_factor.

Value

An object of class "ridge_path", a named list containing:

coef: Numeric array of shape [p, q, length(lambdas)]; coefficient matrix for each penalty.
lambdas: The input penalty vector.
svd: The amSVD factor object used internally.
k: Integer number of singular values actually used.

See also

ridge_fit, svd_factor

Examples

X <- matrix(rnorm(60), nrow = 15)
y <- rnorm(15)
path <- ridge_path(X, y, lambdas = c(0.1, 1, 10))
dim(path$coef)
#> [1] 4 1 3