摘要
基于Zellner的平衡损失的思想,本文提出了矩阵形式的平衡损失函数,并在该损失函数下讨论了多元回归系数线性估计的可容许性.给出了六种不同形式的可容许定义,证明了这六种容许性在齐次和非齐次线性估计类中是一致的,且得到了其共同的可容许估计的充要条件.
In this paper the matrix balanced loss functions are proposed according to Zellner's thought of balanced loss. Admissible linear estimators for regression coefficients in a multivariate linear model are investigated under the matrix balanced loss function under six kinds of definitions for admissibility. The six kinds of admissibility are proved to be identical in a class of homogeneous and nonhomogeneous linear estimators, and the sufficient and necessary conditions that linear estimators are admissible are obtained.
出处
《应用数学学报》
CSCD
北大核心
2009年第5期951-957,共7页
Acta Mathematicae Applicatae Sinica
基金
安徽省高等学校优秀青年人才基金项目(2009SQRZ154)
关键词
多元线性模型
矩阵平衡损失函数
线性估计
可容许性
multivariate linear model
matrix balanced loss function
linear estimators
admissibility