Skip to contents

Compute Weights Using L0Learn

Source: R/regularized_regression.R
l0learn_weights.Rd

Fits an L0-regularized linear regression model via `L0Learn::L0Learn.cvfit` and returns the coefficient vector at the (lambda, gamma) pair minimizing the cross-validation error. Default penalty is "L0"; the user can switch to "L0L1" or "L0L2" (and tune the corresponding gamma grid) by passing the relevant arguments through `...`.

Usage

l0learn_weights(X, y, penalty = "L0", nFolds = 5, ...)

Arguments

X

A numeric matrix of predictors.

y

A numeric response vector.

penalty

Type of regularization: "L0", "L0L1", or "L0L2". Default is "L0".

nFolds

Number of cross-validation folds. Default is 5.

...

Additional arguments passed through to `L0Learn::L0Learn.cvfit` (e.g. `nGamma`, `gammaMin`, `gammaMax`, `algorithm`, `maxSuppSize`).

Value

A numeric vector of length `ncol(X)` of variant weights.