D Lamnisos, JE Griffin and MFJ Steel
Adaptive Monte Carlo for Binary Regression with Many Regressors
Abstract: This article describes a method for efficient posterior simulation for Bayesian variable selection in probit regression models with many regressors but few observations. A proposal on model space is described which contains a tuneable parameter. An adaptive approach to choosing this tuning parameter is described which allows automatic, efficient computation in these models. The methods is applied to the analysis of gene expression data.