deCorr

the preset correlation threshold. Variables with correlation higher than rho will be separate into exclusive groups. Default is set to 0.7

the number of blocks to separate variables

ngrp

Divide the highly correlated variables into exclusive groups

The Proportional Subdistribution Hazard (PSH) model has been popular for estimating the effects of the covariates on the cause of interest in Competing Risks analysis. The fast accumulation of large scale datasets has posed a challenge to classical statistical methods. Current penalized variable selection methods show unsatisfactory performance in ultra-high dimensional data. We propose a novel method, the Random Approximate Elastic Net (RAEN), with a robust and generalized solution to the variable selection problem for the PSH model. Our method shows improved sensitivity for variable selection compared with current methods.

deCorr: De-correlating variables

Description

Usage

Arguments

Value