Bayesian Analysis of Computer Experiments with an Application in Cosmology

• Activity Number: 398
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Sivaranjani Vaidyanathan*+ and Mark Berliner
• Companies: Ohio State University and Ohio State University
• Keywords: Calibration ; Emulators ; Computer experiments ; Gaussian process ; Uncertainty quantificaiton ; Discrepancy

Abstract

We consider the problem of combining computer model output and using field observations to make predictions of the process of interest. We are particularly interested in experiments with highly multivariate responses. We use a Bayesian approach to solve the problem and quantify the resulting uncertainty in parameter estimates and predictions. We apply our method to the problem of constraining cosmological parameters from an observed mass power spectrum and computer simulated data. We present results obtained from using our model and compare that with other approaches that can be used in this setting.