矩阵损失下一般Gauss-Markov模型线性预测的可容许性  

ADMISSIBILITY OF LINEAR PREDICTOR IN THE GENERAL GAUSS-MARKOV MODEL UNDER MATRIX LOSS

作　　者：刘郁文[1] 喻胜华[2] 

机构地区：[1]湖南农业大学理学院,湖南长沙410128 [2]中南大学数学科学与计算技术学院,湖南长沙410083

出　　处：《数学杂志》2007年第2期165-172,共8页Journal of Mathematics

摘　　要：本文研究了一般Gauss-Markov模型中线性可预测变量的线性预测的可容许性.在给出线性预测可容许性定义的基础上,通过构造一个特殊的常量矩阵D0,分别得到了齐次和非齐次线性预测类中可容许的充要条件.In this paper, we investigate the adimissibility of linear prediction of the linear predictable variable in the general Gauss-Markov model. Under the base that the definition of the adimissible linear predictor is given, a special costant matrix Do is constructed, and the necessary and sufficient conditions/or a linear predictor of the linear predictable variable to be adimissble in the class of homogeneous linear predictors and the class nonhomogeneous linear predictors are established respectively.

关 键 词：一般GAUSS-MARKOV模型 线性可预测变量 线性预测 可容许性 

分 类 号：O212.4[理学—概率论与数理统计]