Abstract
We give bounds on the concentration of (pseudo) posterior distributions, both for correct and misspecified models. The bounds are derived using the information inequality, entropy estimates, and empirical process methods.
Information
Published: 1 January 2013
First available in Project Euclid: 8 March 2013
zbMATH: 1353.62029
MathSciNet: MR3202636
Digital Object Identifier: 10.1214/12-IMSCOLL916
Subjects:
Primary:
60K35
Keywords:
posterior contraction, prior, Bayes
Rights: Copyright © 2010, Institute of Mathematical Statistics
