TDfdr: Local False Discovery Rate Estimation with Competition-Based Procedures

Introduction

TDfdr is a local false discovery rate (fdr) estimation method mainly used in the contexts of multiple hypothesis testing and variable selection. Starting from the competition-based procedure, TDfdr takes the advantage of the competitive variables to estimate the null proportion and the null distribution. Then leveraging the iteration framework of kerfdr, TDfdr estimates the fdr semi-parametrically.

R codes

The R code for applying TDfdr can be downloaded here.
The zip file contains the codes that can be used for different scenarios (normal, gamma, and regression data).

 

Contact Us

Any problem with TDfdr please contact:

Yan Fu: yfu(at)amss(dot)ac(dot)cn
Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100049, China.