带约束的增长曲线模型中线性预测的泛容许性  

General Admissibility of Linear Predictor in Growth Curve Model Subject to a Restricted Parameter Set

作　　者：朱宁[1] 周志龙[1] 

机构地区：[1]桂林电子科技大学数学与计算科学学院,广西桂林541004

出　　处：《西南大学学报（自然科学版）》2010年第1期15-20,共6页Journal of Southwest University(Natural Science Edition)

基　　金：广西区"十一五"教改工程项目(GX06066)

摘　　要：在二次损失下研究了带不完全椭球约束的增长曲线模型中未来观察值线性预测的泛容许性.根据线性可预测变量泛容许预测的定义和Φ函数的合理条件,得到了齐次和非齐次线性预测在对应的线性预测类中是泛容许预测的充要条件.This article considers the general admissibility of the linear predictor of future observations in a growth curve model subject to an incomplete ellipsoidal restriction under quadratic loss function.According to the definition for general admissible predictor of a linear predictable variable and the reasonable conditions of function,the necessary and sufficient conditions for a linear predictable variable to be general admissible in classes of the homogeneous or inhomogeneous linear predictors are obtained.

关 键 词：泛容许性 线性预测 增长曲线模型 损失函数 

分 类 号：O212.1[理学—概率论与数理统计]