Bayesian Cholesky Factor Models for Spatial Data

• Activity Number: 608
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Joon Jin Song*+ and Keunbaik Lee
• Companies: Baylor University and Sungkyunkwan University
• Keywords: Spatial Data ; Generalized Linear Mixed Models ; Covariance Estimation ; Modified Cholesky Decomposition

Abstract

It is an important task to estimate covariance in spatial analysis. A major challenge in the estimation is the positive-definiteness constraint. Several approaches have been proposed to provide more accurate and well-conditioned covariance estimators. This work proposes a data-based framework for modeling spatial covariance using the modified Cholesky decomposition that factors covariance matrix into two components modeling variance and dependence. The covariance modeling is embedded in generalized linear mixed models for spatial data. A real data set is used to illustrate the proposed methods.