Rcpp-based objective function for L1- or L2-regularized logistic regression
objective_cpp(beta0, beta, x, y, lambda, alpha = 0)
beta0 | intercept |
---|---|
beta | a feature matrix |
x | observations |
y | response |
lambda | penalty |
alpha | elasticnet mixing parameter |
Objective function value.
if (requireNamespace("glmnet")) { x <- matrix(rnorm(100*20), 100, 20) y <- sample(1:2, 100, replace = TRUE) fit <- glmnet::cv.glmnet(x, y, family = "binomial") lambda <- fit$lambda.1se beta0 <- coef(fit, lambda)[1] beta <- coef(fit, lambda)[-1] objective_cpp(beta0, beta, t(x), y, lambda, alpha = 1) }#> [1] 0.8888095