Holmes, C.; Knorr-Held, Leonhard
(2003):
Efficient simulation of Bayesian logistic regression models.
Collaborative Research Center 386, Discussion Paper 306
|
![[img]](https://epub.ub.uni-muenchen.de/1688/1.hassmallThumbnailVersion/paper_306.pdf)  Preview |
|
315kB |
Abstract
In this paper we highlight a data augmentation approach to inference in the Bayesian logistic regression model. We demonstrate that the resulting conditional likelihood of the regression coefficients is multivariate normal, equivalent to a standard Bayesian linear regression, which allows for efficient simulation using a block Gibbs sampler. We illustrate that the method is particularly suited to problems in covariate set uncertainty and random effects models.