|
Activity Number:
|
453
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #301723 |
|
Title:
|
Assessing Local Model Fit in Bayesian Regression Models Using the Partitioned Deviance Information Criterion
|
|
Author(s):
|
David Wheeler*+ and Lance Waller and DeMarc A. Hickson
|
|
Companies:
|
Emory University and Emory University and University of Mississippi Medical Center
|
|
Address:
|
1518 Clifton Rd, NE Third Floor, Atlanta, GA, 303022,
|
|
Keywords:
|
Bayesian statistics ; spatial statistics ; DIC ; diagnostics ; HIV ; Rwanda
|
|
Abstract:
|
There has been a recent emphasis in the applied spatial statistical literature on models with local covariate effects, instead of more traditional models that represent relationships with static effects across a study area. However, spatially varying coefficient models can be computationally demanding to fit, and diagnostic tools that justify the additional computation effort are welcome. We use a partitioning of the deviance information criterion (DIC) to assess local model fit and observation influence in a Bayesian framework. We map local DIC values and differences in local DIC between models to assist in model selection and to visualize the impact of adding covariates or model parameters. We demonstrate the utility of the local diagnostics with an example of HIV prevalence among pregnant women in the Butare province of Rwanda during 1989-1993 using a range of model specifications.
|
