matrix.convex_relu: Creates a feature matrix for the convex relu problem.

In adelie: Group Lasso and Elastic Net Solver for Generalized Linear Models

Creates a feature matrix for the convex relu problem.

Description

Creates a feature matrix for the convex relu problem.

Usage

matrix.convex_relu(mat, mask, gated = FALSE, n_threads = 1)

Arguments

mat	Base feature matrix. It is either a dense or sparse matrix.
mask	Boolean mask matrix.
gated	Flag to indicate whether to use the convex gated relu feature matrix.
n_threads	Number of threads.

Value

Convex relu feature matrix. The object is an S4 class with methods for efficient computation in C++ by adelie.

Author(s)

Trevor Hastie and James Yang
Maintainer: Trevor Hastie [e-mail]

Examples

n <- 100
p <- 20
m <- 10
Z_dense <- matrix(rnorm(n * p), n, p)
mask <- matrix(rbinom(n * m, 1, 0.5), n, m)
out <- matrix.convex_relu(Z_dense, mask)
Z_sparse <- as(Z_dense, "dgCMatrix")
out <- matrix.convex_relu(Z_sparse, mask)

adelie documentation built on April 3, 2025, 8:58 p.m.