Bayesian shrinkage in mixture of experts models: Identifying robust determinants of class membership

TL;DR

This paper introduces a Bayesian variable selection method for mixture of experts models, using a normal gamma prior, to identify key predictors of class membership, demonstrated on HIV information sources among women in Mozambique.

Contribution

It proposes a novel prior structure for implicit variable selection in mixture of experts models, enhancing robustness in identifying determinants of class membership.

Findings

Identified robust predictors of HIV information sources.

Demonstrated effective Bayesian inference with Gibbs sampling.

Clustered women based on homogeneous HIV information sources.

Abstract

A method for implicit variable selection in mixture of experts frameworks is proposed. We introduce a prior structure where information is taken from a set of independent covariates. Robust class membership predictors are identified using a normal gamma prior. The resulting model setup is used in a finite mixture of Bernoulli distributions to find homogenous clusters of women in Mozambique based on their information sources on HIV. Fully Bayesian inference is carried out via the implementation of a Gibbs sampler.