|
Activity Number:
|
325
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #306500 |
|
Title:
|
Multivariate Spatial Modeling in Bayesian Hierarchical Settings
|
|
Author(s):
|
Sudipto Banerjee*+
|
|
Companies:
|
University of Minnesota
|
|
Address:
|
A460 Mayo Building, MMC 303, Minneapolis, MN, 55455,
|
|
Keywords:
|
geostatistics ; Bayesian modeling ; multivariate spatial data ; cross covariances
|
|
Abstract:
|
Recent advances in Geographical Information Systems (GIS) and the availability of easily accessible databases and software enable statisticians and data analysts to experiment with richer models that capture spatial association in geocoded data. Recently, there has been much interest in the analysis of spatial data concerning multiple variables that arise from different data sources. This talk will present a flexible class of models that arise as space-varying linear transformations of tractable spatial processes. These models yield valid probability models that assign to each variable its own spatial structure. We discuss these methods in the contexts of analyzing multiple cancers over counties in Minnesota with each cancer type having its own spatial structure and for agronomy experiments where spatial association is deemed present at different resolutions.
|
