April 8, 2014
Dr. Xinyu Zhang
Visiting Scholar, Department of Statistics, Texas A&M University
Title: An Overview of Frequentist Model Averaging
Abstract:
Model selection forces us to “putting all our inferential eggs in one unevenly woven basket” (Longford, 2005). Model averaging is an alternative to model selection and often reduces the risk in prediction. In this talk, I will introduce some model averaging methods, including least square model averaging, model averaging for linear mixed effects models, model averaging for high dimensional data, and adaptive model averaging. Asymptotic distribution and mean squared error comparison will also be discussed in this talk.