|
IMPROVED ESTIMATORS OF VARIANCE COMPONENTS
Song Gui WANG, Yong Xu DENG
Acta Mathematicae Applicatae Sinica
1999, 22 (1):
115-122.
DOI: 10.12387/C1999026
Improved estimation of variance components in a linear model with two variance components is considered. For one variance component (say, #sigma#~2>=0) corresponding to a random effect, we study a class of invariant estimator which contains several important estimators in one-way random model. It is proved that there is no best estimator in the class under mean-squared-error criterion and the Analysis-of-variance(ANOVA) estimator is also inadmissible. In an important subclass, however, the best estimator is obtained. for the general case of the model, a class of nonnegative estimators of #sigma#~2 is discussed. The distribution functions of the estimators and a sufficient condition for the superiority of the estimators in the class over the ANOVA estimator are given.
Related Articles |
Metrics
|
|