CIRJE-F-614	"Consistency of the Empirical Bayes Information Criterion for Selecting Variables in Linear Mixed Models"
Author Name	Kubokawa, Tatsuya and Muni S. Srivastava
Date	February 2009
Full Paper
Remarks	Revised version of CIRJE-F-584 (2008); subsequently published in Journal of the Japan Statistical Society 40, No.1, 111-130, 2010. The title of this paper is changed as "An empirical Bayes information criterion for selecting variables in linear mixed models".

Abstract

The paper addresses the problem of selecting variables in the two-stage sampling models characterized as a linear mixed model. We obtain the Empirical Bayes Information Criterion (EBIC) using a prior distribution on regression coefficients with an unknown hyper-parameter. It is shown that EBIC not only has the nice asymptotic property of the consistency as a variable selection, but also performs better in small sample sizes than the conventional methods like BIC and AIC in light of selecting the true variables.