多元回归系数非齐次线性估计在新的优良性准则下的容许性  

Admissibility of Nonhomogeous Linear Estimates of Multivariate Regression Coefficient on a New superior Standard Under the Matrix Loss Function

作　　者：伍长春 

出　　处：《郴州师专学报》1997年第1期56-59,共4页

摘　　要：考虑多线性模型和矩阵损失函数:其中X和U≥0和巳知矩阵,(?)和V≥0(但V≠0)是未知参数矩阵,本文在新的容许性意义下,得到了S(?)的非齐次线性估计LY+a在非齐次线性估计类￡中是可容许的充要条件。Consider the multivariate linear model y =XΘ + ε, E(ε) = 0,Cov(ε) = V U and the matrixLoss function (D-S Θ)(Ⅰ)-S Θ)1,whereX,Sand U≥0 are known matrices,Θand V≥0(but V≠0) are un-known matrices. In this paper. We obtained a necessary and sufficient condition for LY+a to be an admissible estimate of S Θ in the class of nonhomogeous linear estimates and under a new admissible definition.

关 键 词：回归系数 非齐次线性估计 优良性准则 可容许性 

分 类 号：O212.1[理学—概率论与数理统计]