任意秩多元线性模型中最优预测的稳健性  

Robustness of Optimal Prediction in the Multivariate Linear Model with Arbitrary Rank

作　　者：喻胜华[1] 袁权龙[1] 

机构地区：[1]湖南大学数学与计量经济学院,湖南长沙410082

出　　处：《湖南大学学报（自然科学版）》2005年第3期115-118,共4页Journal of Hunan University:Natural Sciences

基　　金：国家自然科学基金资助项目(10101006)

摘　　要：研究了任意秩多元线性模型中最优线性无偏预测的稳健性,即对任一线性可预测变量,得到了其关于协方差矩阵具有稳健性的充要条件.<Abstrcat>Based on the multivariate linear model with arbitrary rank, the unknown observation matrix was predicted by using the known observation matrix. The robustness of prediction was that when the model's covariance matrices had slight perturbation, the previous prediction remained optimal. The robustness of the best linear unbiased predictor in the multivariate linear model with arbitrary rank were investigated.For arbitrary linear predictable variable, the necessary and sufficient condition for it to be robust with respect to covariance matrices was obtained.

关 键 词：多元线性模型 线性可预测变量 最优线性无偏预测 稳健性 

分 类 号：O212.4[理学—概率论与数理统计]