School Seminars and Colloquia

"Gibbs sampling method for linear model selection"

Honours Project Seminar

by Sen Tan

Institution: Department of Mathematics and Statistics, University of Melbourne
Date: Fri 23rd June 2006
Time: 11:30 AM
Location: Russell Love Theatre, Richard Berry Building, The University of Melbourne

Abstract: In modeling a linear model, when the number of potential explanatory variables increases, the number of candidate models increases exponentially. How to overcome this computation difficulty is essential for linear model selection.
In this paper I use a MCMC method---Gibbs sampling method to implement two commonly used model selection criteria AIC and BIC for linear regression model selection where very large number of candidate models are involved, and compare it with other commonly used algorithm such as stepwise and leap and bounds method.

For More Information: Associate Professor Felisa J. Vázquez-Abad