complexity.LASSO: Complexity selection helpers for penalized backends

In peperr: Parallelised Estimation of Prediction Error

Complexity selection helpers for penalized backends

Description

Wrapper functions for choosing the penalized tuning parameter used by fit.penalized, fit.LASSO, and fit.fusedLASSO.

Usage

complexity.cv.penalized(response, x, full.data, ...)
complexity.LASSO(response, x, full.data, ...)
complexity.fusedLASSO(response, x, full.data, ...)

Arguments

response	response used for model fitting. Survival responses may be passed as a Surv object or as a two-column matrix with time and status.
x	n * p matrix of covariates.
full.data	data frame containing the full data set. This argument is accepted to match the peperr complexity-function contract.
...	additional arguments passed to profL1.

Details

complexity.cv.penalized profiles the cross-validated lambda1 values returned by profL1 and keeps the maximizer. complexity.LASSO is a compatibility alias. complexity.fusedLASSO performs the same search while forwarding fusedl = TRUE.

Value

A scalar tuning value, typically the selected lambda1.

See Also

fit.penalized, fit.LASSO, fit.fusedLASSO, profL1

peperr documentation built on March 25, 2026, 9:06 a.m.