Abstract
LetY have ann-variate normal distribution with covariance matrixσ 2I and mean vectorXβ, whereX is a knownn×p matrix. The problem of estimatingθ=σ 2+β′X′CXβ is studied. The admissibility and inadmissibility of the estimators of the form\(b\hat S^2 + \hat \beta 'X'CX\hat \beta \), where\(\hat \beta = (X'X)^ - X'Y\) and\(S^2 = (Y - X\hat \beta )'(Y - X\hat \beta )\), are established. Another class of admissible quadratic estimators ofθ is derived.
Similar content being viewed by others
References
Cheng P. Admissibility of Simultaneous Estimation of Several Parameters.J. Sys. Sci. & Math. Sci., 1982, 2: 176–195.
Cheng P., Wu Q.G. and Li G.Y. Admissibility of Quadratic Estimates of 2-order Moment About the Origin.Acta Math. Appl. Sinica, 1983, 6: 18–28 (in Chinese).
Rukhin, A.L. Quadratic Estimators of Quadratic Functions of Normal Parameters.Journal of Statistical Planning and Inference, 1987, 15: 301–310.
Stein, C. Inadmissibility of the Usual Estimator of the Variance of a Normal Distribution with Unknown Mean.Ann. Inst. Statist. Math., 1964, 16: 155–160.
Zidek, J.V. Sufficient Condition for the Admissibility Under Squared Error Loss for Formal Bayes Estimators.Ann. Math. Statist., 1970, 41: 446–465.
Author information
Authors and Affiliations
Additional information
This work is supported by the National Natural Science Foundation of China.
Rights and permissions
About this article
Cite this article
Wu, Q. Quadratic estimators of quadratic functions with parameters in normal linear models. Acta Mathematicae Applicatae Sinica 11, 378–388 (1995). https://doi.org/10.1007/BF02007176
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02007176