run_BHq_white_noise: Run Benjamini-Hochberg procedure with added noise

In svteichman/prelim.knockoffs: Implements the knockoff method by Barber & Candes

Description

Runs the altered Benjamini-Hochberg procedure, which adds a mean zero normal RV to each beta hat to make the z-scores independent of each other.

Usage

run_BHq_white_noise(X, y, fdr)

Arguments

X	A nxp design matrix.
y	A vector of n responses.
fdr	The desired false discovery rate.

Value

A list containing:

X	n-by-p design matrix.
y	vector of observed responses.
Z	vector of test statistics.
thres	data-dependent threshold.
selected	list of selected variables.

svteichman/prelim.knockoffs documentation built on May 28, 2020, 5:14 p.m.