|
Activity Number:
|
137
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #309426 |
|
Title:
|
Objective Priors for Overdisposed Weibull Models for Breast Cancer Incidence and Survivals
|
|
Author(s):
|
Luyan Dai*+ and Zhuoqiong (Chong) He and Dongchu Sun and Mario Schootman
|
|
Companies:
|
University of Missouri-Columbia and University of Missouri-Columbia and University of Missouri-Columbia and Washington University in St. Louis
|
|
Address:
|
29 Broadway Village Dr, Columbia, MO, 65201,
|
|
Keywords:
|
Refernce prior ; spatial-survival model ; posterior quantiles ; objective priors ; breast cancer
|
|
Abstract:
|
Most of the widely used Weibull models in Bayesian survival analysis are based on subjective priors. The Bayesian estimates are often very sensitive to the choices of the priors. Some known objective priors for overdisposed Weibull models would result in improper posteriors. In this paper, we propose a spatial-survival model by incorporating reference priors to investigate geological variation on county and census tract level among breast cancer incidents in Iowa. Moreover, the estimators of unknown parameters are computed by posterior quantiles since reference priors lead to nonexistence of posterior moments in this case. Results based on subjective and objective priors are also compared.
|
