Faculty Advisor

Dr. Steven Carr

Department

Computer Science

Presentation Date

4-14-2016

Document Type

Poster

Abstract

We apply high performance numerical integration to problems in Bayesian statistics. These are applied to data arising in the analysis of problems in such areas as medical statistics (birth weight data, heart transplant data, photocarcinogen data, radio therapy data), climate (global weather, tornado data), and general statistics applications (multivariate logistic distribution, multivariate normal, nonlinear regression). We compute Bayesian moment integrals using the ParInt integration software that runs efficiently on computer clusters. We compare our results to those in the literature and show excellent performance with respect to accuracy and execution time.

Share

COinS