Title: Controlling false discovery rate using Gaussian mirrors
Authors: Xin Xing - Virginia Tech University (United States)
Zhigen Zhao - Temple University (United States) [presenting]
Jun Liu - Havard Univeristy (United States)
Abstract: Simultaneously finding multiple influential variables and controlling the false discovery rate (FDR) for linear regression models is a fundamental problem with a long history. We propose the Gaussian Mirror (GM) method, which creates for each predictor variable a pair of mirror variables by adding and subtracting a randomly generated Gaussian random variable and proceeds with a certain regression method, such as the ordinary least-square or the Lasso. The mirror variables naturally lead to a test statistic highly effective for controlling the FDR. Under a weak dependence assumption, we show that the FDR can be controlled at a user-specified level asymptotically. It is shown that the GM method is more powerful than many existing methods in selecting important variables, subject to the control of FDR, especially under the case when high correlations among the covariates exist.