generate.knockoff: Generate Knockoff Matrix

In KOBT: Knockoff Boosted Tree

Description

Generate different types of knockoff matrices given an original one.

Usage

generate.knockoff(X, type, num, num.comp = 10)

Arguments

X	An input original design matrix.
type	The knockoff type to be generated. There are three choices available: (1) "shrink" for the shrink Gaussian knockoff; (2) "sparse" for the sparse Gaussian knockoff; and (3) "pc" for the pricial component knockoff.
num	The number of knockoff matrices to be created.
num.comp	The number of pricial components to be used for generating knockoff matrices, the default is 10.

Value

A list of created knockoff matrices.

References

\insertRef

barber2015controllingKOBT \insertRefcandes2018panningKOBT \insertRefbien2011sparseKOBT

Examples

set.seed(10)
X <- matrix(rnorm(100), nrow = 10)
Z <- generate.knockoff(X = X, type = "shrink", num = 5)

KOBT documentation built on Feb. 20, 2020, 5:08 p.m.