N. P. Baldin, V. G. Spokoiny, “Bayesian model selection and the concentration of the posterior of hyperparameters”, Fundam. Prikl. Mat., 2013, Volume 18, Issue 2,Pages <nobr>13

Bayesian model selection and the concentration of the posterior of hyperparameters

N. P. Baldin^a, V. G. Spokoiny^abc

^a Laboratory of Structural Methods of Data Analysis in Predicative Modeling, Moscow Institute of Physics and Technology, Moscow, Russia
^b Weierstrass Institute for Applied Analysis and Stochastics, Berlin, Germany
^c Humboldt University of Berlin, Berlin, Germany

Abstract: The present paper offers a construction of a hyperprior that can be used for Bayesian model selection. This construction is inspired by the idea of the unbiased model selection in a penalized maximum likelihood approach. The main result shows a one-sided contraction of the posterior: the posterior mass is allocated on models of lower complexity than the oracle one.

UDC: 519.22