Importance sampling methods for Bayesian discrimination between embedded models

Abstract

This paper surveys some well-established approaches on the approximation of Bayes factors used in Bayesian model choice, mostly as covered in Chen et al. (2000). Our focus here is on methods that are based on importance sampling strategies rather than variable dimension techniques like reversible jump MCMC, including: crude Monte Carlo, maximum likelihood based importance sampling, bridge and harmonic mean sampling, as well as Chib's method based on the exploitation of a functional equality. We demonstrate in this survey how these different methods can be efficiently implemented for testing the significance of a predictive variable in a probit model. Finally, we compare their performances on a real dataset.

References in corpus (2)

• Properties of Nested Sampling
• Computational methods for Bayesian model choice

Cited by in corpus (15)

• Testing hypotheses via a mixture estimation model
• On resolving the Savage-Dickey paradox
• On Particle Learning
• The Controlled Thermodynamic Integral for Bayesian Model Comparison
• Partition Functions from Rao-Blackwellized Tempered Sampling
• A Parsimonious Tour of Bayesian Model Uncertainty
• The expected demise of the Bayes factor
• Rao-Blackwellization in the MCMC era
• Discussions on "Riemann manifold Langevin and Hamiltonian Monte Carlo methods"
• Importance sampling schemes for evidence approximation in mixture models
• Variational Bayes approach for model aggregation in unsupervised classification with Markovian dependency
• Bayesian Computational Tools
• Using parallel computation to improve Independent Metropolis--Hastings based estimation
• On computational tools for Bayesian data analysis
• On Bayesian Data Analysis