带线性不等式约束的线性模型中线性预测的可容许性  

Admissibility of Linear Prediction in the Linear Model with Linear Inequality Constraints

作　　者：王浩波[1] 罗建华[1] 

机构地区：[1]中南林业科技大学理学院,长沙410004

出　　处：《工程数学学报》2010年第3期441-445,共5页Chinese Journal of Engineering Mathematics

基　　金：中南林业科技大学人才引进基金(101-0638)~~

摘　　要：本文刻画了线性模型在线性不等式约束条件下的线性预测的可容许性。我们给出了条件线性可预测变量和线性不等式约束条件下可容许预测的定义,在二次损失函数下,讨论了齐次和非齐次线性预测可容许性的关系,得到了条件线性可预测变量的线性预测是可容许线性预测的充要条件。The admissibility of linear predictions in linear models is characterized in this paper. We give the definitions of the conditional linear predictable variable and the admissible prediction with linear inequality constraints. Under the quadratic loss function, the relation between admissibility of inhomogeneous and homogeneous linear predictions is discussed, and some necessary and sufficient conditions for a linear prediction of conditional linear predictable variables to be admissible are obtained.

关 键 词：二次损失 线性不等式约束 条件线性可预测变量 可容许性 

分 类 号：O212.1[理学—概率论与数理统计]