Admissibility of linear estimators of the common mean parameter in general linear models under a balanced loss function

详细信息    查看全文

• 作者：Ming-Xiang Cao^a ; ^b ; ^{caomingx@163.com} ; Dao-Jiang He^a
• 关键词：primary ; 62C15 ; secondary ; 62J05
• 刊名：Journal of Multivariate Analysis
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：153
• 期：Complete
• 页码：246-254
• 全文大小：529 K
• 卷排序：153

文摘

In order to investigate linearly admissible estimators of the common mean parameter in general linear models, we introduce and motivate the use of a balanced loss function obtained by combining Zellner’s idea of balanced loss (Zellner, 1994) with the unified theory of least squares (Rao, 1973). In classes of homogeneous and non-homogeneous linear estimators, sufficient and necessary conditions for linear estimators of the common mean parameter to be admissible are obtained, respectively. A comparison is then made between linearly admissible estimators and a “truly” unified least square estimator.