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 - #301823 |
Title:
|
A Model Search Strategy for Probit Models with Binary Covariates
|
Author(s):
|
German Molina*+ and James Berger+
|
Affiliation(s):
|
Duke University and Duke University
|
Address:
|
222 old chem. Building, Durham, North Carolina, 27708, USA Box 90251, Durham, North Carolina, 27708,
|
Keywords:
|
Bayesian ; Model space search ; Probit ; Binary data ; Model averaging
|
Abstract:
|
Our motivating example requires probit analysis of binary data. We have 7 binary variables of interest. These define 128 possible covariates, out of which 72 are possible. This leaves us with 2^72 possible models to explore (some less since we restrict ourselves to nested models). The target is to explore this model space in a fast and computationally efficient form. We define the noninformative prior on the full model and induce by either marginalizing or conditioning the corresponding priors on the parameters of the submodels. The Laplace approximation is used to compute the posterior probability of each such submodel. We also present a possible scheme, which makes use of the posteriors previously computed, for searching through the model space. Median probability model and model averaging are used for inference on the parameters of interest. A comparison with the BIC-based results is provided.
|