Superiority of empirical Bayes estimation of error variance in linear model

Abstract

In this paper, the Bayes estimator of the error variance is derived in a linear regression model, and the parametric empirical Bayes estimator (PEBE) is constructed. The superiority of the PEBE over the least squares estimator (LSE) is investigated under the mean square error (MSE) criterion. Finally, some simulation results for the PEBE are obtained.

Keywords

Linear regression model / error variance / parametric empirical Bayes estimation / mean square error criterion / simulation result

References

[1] Berger J. O. Statistical Decision Theory and Bayesian Analysis, 1985, Berlin: Springer

[2] Box G. P., Tiao G. C. Bayesian Inference in Statistical Analysis, 1973, Reading: Addison-Wesley

[3] Broemeling L. D. Bayesian Analysis of Linear Models, 1985, New York: Marcel Dekker

[4] Chen Ling. Studies of some issues of the parametric empirical Bayes estimation in linear model. Doctoral Thesis 2011-JX-C-250, 2011, Hefei: University of Science and Technology of China

[5] Efron B., Morris C. Empirical Bayes on vector observation: an extension of stein’s method. Biometrika, 1972, 59: 335-347

[6] Ghosh M., Saleh A. K., Md E., Sen P. K. Empirical Bayes subset estimation in regression models. Statist Decisions, 1989, 7: 15-35

[7] Kotz S., Nadarajah S. Multivariate t Distributions and Their Applications, 2004, Cambridge: Cambridge University Press

[8] Morris C. Parametric empirical Bayes inference: theory and applications. J Amer Statist Assoc, 1983, 78: 47-65

[9] Robbins H. An empirical Bayes approach to statistics. In: Proc Third Berkeley Symp on Math Statist Prob, Vol 1, 1956, Berkeley: Univ of California Press, 157-163

[10] Singh R. S. Empirical Bayes estimation in a multiple linear regression model. Ann Inst Statist Math, 1985, 37: 71-86

[11] Wang S G, Chow S C. Advanced Linear Models: Theory and Applications. New York: Marcel Dekker, 1994

[12] Wei L. S., Chen J. H. Empirical Bayes estimation and its superiority for two-way classification model. Statist Probab Lett, 2003, 63: 165-175

[13] Wei L. S., Trenkler G. Mean square error matrix superiority of empirical Bayes estimators under misspecification. Test, 1995, 4: 187-205

[14] Wei L. S., Zhang S. P. The convergence rates of empirical Bayes estimation in a multiple linear regression model. Ann Inst Statist Math, 1995, 47: 81-97

[15] Wei L. S., Zhang W. P. The superiorities of Bayes linear minimum risk estimation in linear model. Comm Statist: Theory Methods, 2007, 36: 917-926

[16] Zhang W. P., Wei L. S., Yang Y. N. The superiorities of empirical Bayes estimator of parameters in linear model. Statist Probab Lett, 2005, 72: 43-50