A Flexible Hierarchical Approach to Modeling Discrete-Valued Spatio-Temporal Data

• Activity Number: 317
• Type: Invited
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Scott H. Holan*+ and Christopher K. Wikle
• Companies: University of Missouri and University of Missouri
• Keywords: Bayesian hierarchical models ; counts ; overdispersion ; spatially-varying

Abstract

In many cases modeling discrete-valued spatio-temporal data is a straightforward endeavor. However, in many real-world applications the complexities of the data and/or process don't allow for routine model specification. For example, often discrete-valued spatio-temporal data exhibit zero-inflation, over/under dispersion or heavy tails and contain many sources of uncertainty. In order to accommodate such structure, while quantifying different sources of uncertainty, we propose a hierarchical Bayesian model that utilizes a flexible likelihood specification. The approach we propose allows the likelihood to adapt to the nuances of the discrete-valued data while flexibly accommodating different spatio-temporal dependence structures. The effectiveness of our methodology is demonstrated through simulation and through a real-data application.