Open Access
June 2015 Bayesian Model Selection Based on Proper Scoring Rules
A. Philip Dawid, Monica Musio
Bayesian Anal. 10(2): 479-499 (June 2015). DOI: 10.1214/15-BA942


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.


Download Citation

A. Philip Dawid. Monica Musio. "Bayesian Model Selection Based on Proper Scoring Rules." Bayesian Anal. 10 (2) 479 - 499, June 2015.


Published: June 2015
First available in Project Euclid: 4 February 2015

zbMATH: 1335.62017
MathSciNet: MR3420890
Digital Object Identifier: 10.1214/15-BA942

Keywords: consistent model selection , homogeneous score , Hyvärinen score , prequential

Rights: Copyright © 2015 International Society for Bayesian Analysis

Vol.10 • No. 2 • June 2015
Back to Top