rpc (version 2.0.3)

eBIC: Model selection using extended BIC

Description

This function performs model selection using extended BIC and the ridge partial correlation coefficients

Usage

eBIC(rpc.obj)

Value

BIC.PATH vector of eBIC of each model, starting with the intercept only model.

model.best sorted indices (in increased order) of the model with the smallest EBIC

Arguments

rpc.obj

an object of class rpc, containing the following items:
- X: matrix of covariates (excluding intercept)
- y: response vector
- rpc: vector of ridge partial correlation coefficients, one for each column in X.

Author

An Nguyen

Somak Dutta

Maintainer: Somak Dutta <[e-mail]>

Examples

n <- 50; p <- 400;
trueidx <- 1:3
truebeta <- c(4,5,6)
X <- matrix(rnorm(n*p), n, p) # n x p covariate matrix
y <- 0.5 + X[,trueidx] %*% truebeta + rnorm(n)
res <- rpc(X,y, lambda = 0.1, ncores = 1)
eBIC(res) # model.best: model selected by eBIC