鞅差误差序列下半参数EV回归模型的近邻估计  

THE NEAR NEIGHBOR ESTIMATOR IN SEMI-PARAMETRIC AND ERROR IN VARIABLES REGRESSION MODEL UNDER THE MARTINGALE DIFFERENCE ERROR SEQUENCE CASE

作　　者：谭星[1] 

机构地区：[1]武汉大学数学与统计学院,湖北武汉430072

出　　处：《数学杂志》2008年第2期203-208,共6页Journal of Mathematics

基　　金：国家自然科学基金(10371092)

摘　　要：本文研究了误差为鞅差序列的条件下的一维半参数EV回归模型.利用两步估计的方法构造了参数分量和非参数分量的近邻估计,并且分别证明了估计量的L2相合性和强相合性,从而推广了在普通半参数回归模型已有的相关结论.In this paper, under martingale difference error sequence, the one dimensional semi- parametric and error in variables regression model is studied. By the two steps method, the inadequate near neighbor estimators of both parameter and non-parameter are established. Moreover, the consistency of these estimators are proved respectively, which extends the existed results under the normal semi-parametric regression model.

关 键 词：半参数回归模型 EV回归模型 近邻估计 鞅差序列 相合性 

分 类 号：O212.1[理学—概率论与数理统计]