Activity Number:
|
253
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Bayesian Stat. Sciences*
|
Abstract - #301769 |
Title:
|
A New Simulation Method for Bayesian Model Mixing
|
Author(s):
|
Giovanni Petris*+
|
Affiliation(s):
|
University of Arkansas
|
Address:
|
, Fayetteville, Arkansas, 72701, USA
|
Keywords:
|
MCMC ; Bayesian inference ; Model averaging
|
Abstract:
|
We propose an efficient and easy-to-implement method to sample from a multidimensional distribution with nested degenerate masses. The method is based on a transformation of the mixed distribution into an absolutely continuous one. We show how, when the distribution of interest is the posterior corresponding to a prior with degenerate masses, one can avoid computing the relative weights of the various components (usually performed via numerical integration). Examples of applications are provided.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |