Scalable Bayesian Variable Selection for High Dimensional Data

Naveen Narisetty, University of Illinois Urbana Champagne
Monday, November 27, 2017 - 4:00pm
Middlebush Hall 212

Abstract: We consider the computational and statistical issues for high dimensional Bayesian model selection under the Gaussian spike and slab priors. To avoid large matrix computations needed in standard Gibbs sampling algorithms, we propose a novel Gibbs sampler called "Skinny Gibbs" which is much more scalable to high dimensional problems, both in memory and in computational efficiency. In particular, it's computational complexity grows only linearly in p, the number of predictors, while retaining the property of strong model selection consistency even when p is much greater than the sample size n.