Bayesian Model Selection Based on Proper Scoring Rules

Abstract

Bayesian model selection with improper priors is not well-defined because of the dependence of the marginal likelihood on the arbitrary scaling constants of the within-model prior densities. We show how this problem can be evaded by replacing marginal log-likelihood by a homogeneous proper scoring rule, which is insensitive to the scaling constants. Suitably applied, this will typically enable consistent selection of the true model.

Published at http://dx.doi.org/10.1214/15-BA942 in the Bayesian Analysis (http://projecteuclid.org/euclid.ba) by the International Society of Bayesian Analysis (http://bayesian.org/)

References in corpus (2)

• Criteria for Bayesian model choice with application to variable selection
• Consistency of objective Bayes factors as the model dimension grows

Cited by in corpus (5)

• Objective Bayesian inference with proper scoring rules
• Bayesian model comparison with the Hyvärinen score: computation and consistency
• On Selection Criteria for the Tuning Parameter in Robust Divergence
• A Kernel Stein Test for Comparing Latent Variable Models
• A Note on Bayesian Model Selection for Discrete Data Using Proper Scoring Rules