|
Activity Number:
|
286
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #309296 |
|
Title:
|
Bayesian Analysis of Cross-Classified Spatial Data with Autocorrelation
|
|
Author(s):
|
Xiaolei Li*+ and Murray K. Clayton
|
|
Companies:
|
GlaxoSmithKline and University of Wisconsin-Madison
|
|
Address:
|
812 Morgan Dr, Royeersford, PA, 19468,
|
|
Keywords:
|
MCMC ; Spatial Statistics ; Categorical Data ; Gibbs Sampler ; Cross-classification ; high-dimentional parameter space
|
|
Abstract:
|
The work is focused on the development and application of statistical methodologies to the analyses of categorical data collected over space. When several spatial attributes are considered simultaneously, their mutual associations are hard to characterize. The standard chi-squared analysis becomes invalid and can lead to wrong conclusions because of the spatial autocorrelation within each attribute. Our methods focus on identifying the mutual independence between two multi-categorical spatial processes over a finite lattice. Multinomial autologistic Markov models are constructed for more than one multi-categorical spatial processes as well as the mutual dependence between any two of them. For model inferences, a new MCMC algorithm are proposed for estimations in high-dimensional parameter space and combined with Gibbs sampler. Then, this Bayesian procedure is justified theoretically.
|
