一个半参数回归模型中Cramér渐近有效性的充要条件(英语)  

A Sufficient and Necessary Condition on Cramer Asymptotic Efficiency in a Semiparametric Regression Model

作　　者：梁华[1] 

机构地区：[1]中国科学院系统科学研究所,北京100080

出　　处：《应用概率统计》1996年第3期233-238,共6页Chinese Journal of Applied Probability and Statistics

基　　金：Supported by RQ9201 Foundation

摘　　要：本文考虑模型Y_i=X^Tβ+g(T)+ε,这里(X^T,T,Y,)是k+2-维随机向量,g是未知光滑函数,ε是均值为零方差有限的随机误差。本文证明了,的最小二乘估计是Cramer渐近有效的充要条件是误差6服从正态分布N(0,口’)。The paper is concerned with the semiparametric regression model Y = Xrβ+ g(T) +ε, where (Xr ,T, Y) are (k + 2)-dimensional random vector, g(·) is an unknown smooth function to be estimated, ε is variable error with zero mean and finite variance It is shown that the LS estimators of β is asymptotically efficiency in sense of Cramer if and only if ε- N(0, σ2).

关 键 词：半参数回归模型 回归模型 γ渐近有效性 

分 类 号：O212.4[理学—概率论与数理统计]