Bayesian Variable Selection for Median Latent Variable Model

• Activity Number: 130
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Yifan Wang*+ and Zhibin Xu and Hong Ji
• Companies: and Capital University of Economics and Business and Capital University of Economics and Business
• Keywords: median regression ; confirmatory factor analysis model ; normal-gamma prior ; asymmetric Laplace distribution ; Markov chain Monte Carlo

Abstract

In biomedical, psychological, social, and behavioral sciences, it is very common to encounter latent variables along with non-normal data. We propose a median latent variable model to deal with this kind of data in a Bayesian framework. The normal-gamma prior distribution is applied here for simultaneous estimation and model selection. A Markov chain Monte Carlo (MCMC) algorithm for obtaining Bayesian estimates is developed. Simulation studies are carried out to examine the finite sample performance of the proposed estimators. We illustrate the proposed method with a real data set from a longitudinal study of polydrug use.