Bayesian Variable Selection with Dependent Priors for Regularization Parameters

• Activity Number: 244
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Changgee Chang* and Suprateek Kundu and Qi Long
• Companies: Emory University and Emory University and Emory University
• Keywords: Bayesian Regularization ; Variable Selection ; EM algorithm

Abstract

Recently, substantial effort has been made to incorporate biological pathway information among covariates into variable selection. In this work, we adopt a Bayesian regularization approach for variable selection. Different from other approaches, however, our method uses dependent priors for regularization parameters in order to incorporate the pathway information. We develop an EM algorithms with Laplace approximation and present fast computing algorithms. We examine the performance of our proposed method by simulation studies and a real data example.