Tuesday, February 23rd, 2016
2:00pm - 3:00pm, in McCormack 2-116

Li Xiaoou

Columbia University

Level Crossing of Random Functions and Its Applications in Statistics

Abstract: Level crossing of random functions is a classic problem. The literature has been focused on Gaussian random fields. In this talk, I will develop a general framework for analyzing level crossing of possibly non-Gaussian random functions derive asymptotic approximations of the level crossing probabilities. I will also discuss several statistical applications including precise risk bounds for classification, hypothesis testing, signal processing, and sequential analysis.