Abstract

This talk will be an introduction to Stein's method of exchangeable pairs for distributional approximation, assuming no prior knowledge of the method. I'll discuss the ideas behind the method in general, and how it is implemented in order to prove Gaussian limit theorems (with convergence rates) in univariate and multivariate contexts. I'll give an outline of how one can use the method to prove a multivariate central limit theorem, and if there is time, I will briefly discuss how to use the method for Poisson approximation.

MONDAY, February 18, 2008
Time: 5:00 P.M.
Location: WeH 6423