A General Framework for Information Pooling in Two-Sample Multiple Testing

题目：A General Framework for InformationPooling in Two-Sample Multiple Testing

报告人： WenguangSun (University of Southern California)

时间：2018.12.12(周三)下午 2:00

地点：紫金港校区管理学院行政楼14楼1417报告厅

摘要：

In this talk, we discuss ageneral framework for exploiting the sparsity information in two-samplemultiple testing problems. We propose to first construct a covariate sequence,in addition to the usual primary test statistics, to capture the sparsity structure,and then incorporate the auxiliary covariates in inference via a three-stepalgorithm consisting of grouping, adjusting and pooling (GAP). The GAPprocedure provides a simple and effective framework for information pooling. Animportant advantage of GAP is its capability of handling various dependencestructures such as those arise from multiple testing for high-dimensionallinear regression, differential correlation analysis, and differential networkanalysis. We establish general conditions under which GAP is asymptoticallyvalid for false discovery rate control, and show that these conditions arefulfilled in a range of applications. Numerical results demonstrate thatexisting methods can be significantly improved by the proposed framework. This isthe joint work with Yin Xia from Fudan University and Tony Cai from Universityof Pennsylvania.

欢迎广大师生踊跃参加!

联系人：张立新 （stazlx@zju.edu.cn）