BIMSA >
Data Analysis and Problem Solving Seminar
Data Analysis and Problem Solving Seminar
Recent Advances in Multiple Hypothesis Testing with Some Applications
Recent Advances in Multiple Hypothesis Testing with Some Applications
Organizer
Speaker
Kejun He
Time
Friday, May 8, 2026 3:00 PM - 4:00 PM
Venue
A3-1-301
Online
Zoom 204 323 0165
(BIMSA)
Abstract
Unlike traditional single hypothesis testing that assesses one statistical assumption at a time, multiple hypothesis testing conducts numerous tests simultaneously within a single study—often leading to severe inflation of false positive risks. One key objective of multiple hypothesis testing is to detect true signals effectively while limiting the proportion of false positives among all rejections. In this talk, I will present some recent advances in multiple hypothesis testing, such as improving detection power for spatial signals by borrowing neighboring information, identifying the joint statistical significance of individual features, and a series of studies based on the e-value. Instead of delving into methodological details, I will illustrate the value of statistical tools in the real world using examples from environmental sciences, mediation analysis, and industrial predictive maintenance.
Speaker Intro
Dr. He is an Associate Professor at the Institute of Statistics and Big Data, Renmin University of China. He earned his Ph.D. in Statistics from the Department of Statistics at Texas A&M University. His research focuses on tensor data analysis, nonparametric statistics, and functional data analysis. He has authored over 30 papers published in leading journals, including JRSSB, JASA, JMLR, Bernoulli, and Biometrics. He has supervised 6 doctoral students to graduation, whose current positions include assistant professorship at renowned universities in North America and in China.