Two-level stochastic search variable selection in GLMs with missing predictors


Mitra, Robin and Dunson, David (2010) Two-level stochastic search variable selection in GLMs with missing predictors. International Journal of Biostatistics, 6, (1), 33. (doi:10.2202/1557-4679.1173).

Download

Full text not available from this repository.

Description/Abstract

Stochastic search variable selection (SSVS) algorithms provide an appealing and widely used approach for searching for good subsets of predictors while simultaneously estimating posterior model probabilities and model-averaged predictive distributions. This article proposes a two-level generalization of SSVS to account for missing predictors while accommodating uncertainty in the relationships between these predictors. Bayesian approaches for allowing predictors that are missing at random require a model on the joint distribution of the predictors. We show that predictive performance can be improved by allowing uncertainty in the specification of predictor relationships in this model. The methods are illustrated through simulation studies and analysis of an epidemiologic data set.

Item Type: Article
ISSNs: 1557-4679 (print)
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics
Divisions: University Structure - Pre August 2011 > School of Mathematics > Statistics
ePrint ID: 181607
Date Deposited: 19 Apr 2011 10:34
Last Modified: 27 Mar 2014 19:34
URI: http://eprints.soton.ac.uk/id/eprint/181607

Actions (login required)

View Item View Item