Screening in Computer Experiments Using Bayesian Composite Process Models

• Activity Number: 77
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Casey Davis*+ and Christopher Hans and Thomas J. Santner
• Companies: Ohio State University and Ohio State University and Ohio State University
• Keywords: nonstationary ; variable selection ; Bayesian ; Gaussian process ; hierarchical

Abstract

This research develops screening methodology for a computer experiment with many inputs that is based on a hierarchical Bayesian Gaussian process model. The method is based on an extension of the Composite Gaussian Process of Ba and Joseph (2012) and has a non-stationary covariance. The likelihood stage of the interpolating model combines two independent Gaussian processes and the remaining stages put priors on the means, variances, and correlation parameters of the Gaussian processes. This flexible prediction model is able to describe output functions having varying range and patterns of fluctuation. Screening is accomplished by identifying inputs with small posterior probability of being correlated with the output by incorporating a Bayesian "variable selection" prior for the correlation parameters.