Asymptotically Optimal Empirical Bayes Estimation for Variance Components in Random Effects Models
Received:October 14, 2002  
Key Word: Random effects models   variance components   Bayes estimators   empirical Bayes estimators   asymptotic optimality.  
Fund ProjectL:
Author NameAffiliation
WEI Lai-sheng Department of Statistics and Finance
University of Science & Technology of China
Hefei
China 
WANG Li-chun Department of Statistics and Finance
University of Science & Technology of China
Hefei
China 
Hits: 1521
Download times: 860
Abstract:
      In this paper, the Bayes estimators of variance components for two-way classification random effects models is derived under weighted quadratic loss function, and the empirical Bayes (EB) estimators are constructed by the kernel estimation of multivariate density and its mixed partial derivatives. The asymptotically optimality of the EB estimators are obtained under some suitable conditions, and the special cases and the generalizations of the model are shown. Finally, an example satisfying the conditions of theorem is given.
Citation:
DOI:10.3770/j.issn:1000-341X.2004.04.010
View Full Text  View/Add Comment  Download reader