component_information: Information criteria for component selection

In bigPLScox: Partial Least Squares for Cox Models with Big Matrices

Information criteria for component selection

Description

Computes log-likelihood, AIC and BIC values for nested models using the latent components estimated by big_pls_cox() or big_pls_cox_gd().

Usage

component_information(object, max_comp = ncol(object$scores))

## S3 method for class 'big_pls_cox'
component_information(object, max_comp = ncol(object$scores))

## S3 method for class 'big_pls_cox_gd'
component_information(object, max_comp = ncol(object$scores))

select_ncomp(object, criterion = c("AIC", "BIC", "loglik"), ...)

Arguments

object	A fitted object of class big_pls_cox or big_pls_cox_gd.
max_comp	Maximum number of components to consider. Defaults to all components stored in the model.
criterion	Criterion to optimise: "AIC", "BIC" or "loglik".
...	Passed to component_information().

Value

A data frame with columns ncomp, loglik, AIC, and BIC.

A list with the table of information criteria and the recommended number of components.

bigPLScox documentation built on Nov. 18, 2025, 5:06 p.m.