Toward Efficient MCMC for Some High-Dimensional Latent Variable Models

• Activity Number: 463
• Type: Invited
• Date/Time: Wednesday, August 12, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Murali Haran*
• Companies: Penn State
• Keywords: Markov chain Monte Carlo ; latent variables ; composite likelihood ; dimension reduction

Abstract

Among the great successes of Markov chain Monte Carlo (MCMC) methods is their ability to fit latent variable models. In practice, however, if the number of latent variables is large, designing an efficient MCMC algorithm becomes very difficult and MCMC generally becomes prohibitively expensive. I will discuss some approaches for addressing this challenge in latent variable models for spatial data. Among the methods I will describe are reparameterization and dimension-reduction approaches as well as composite likelihood-based methods.