lasso2 (version 1.2-19)

vcov.l1ce: Variance-Covariance Matrix of `l1ce' or `l1celist' Objects

Description

This is a method for the function vcov() for objects inheriting from class l1ce or l1celist (see help(l1ce.object) and help(l1celist.object)). See vcov for the general behavior of this function.

Usage

# S3 method for l1ce
vcov(object, type = c("OPT", "Tibshirani"),
    gen.inverse.diag = 0, …)
# S3 method for l1celist
vcov(object, type = c("OPT", "Tibshirani"),
    gen.inverse.diag = 0, …)

Arguments

object

an object of class l1ce or l1celist.

type

character indicating whether to use the covariance formula of Osborne, Presnell and Turlach or the formula of Tibshirani.

gen.inverse.diag

if Tibshirani's formula for the covariance matrix is used, this value is used for the diagonal elements of the generalised inverse that appears in the formula that corresponds to parameters estimated to be zero. The default is 0, i.e. use the Moore-Penrose inverse. Tibshirani's code uses gen.inverse.diag = 1e11.

further potential arguments passed to methods.