缺失数据下部分线性EV模型的估计及其渐进性质  

Estimation and Asymptotic Property of Partial Linear EV Model with Data Missing

作　　者：夏亚峰[1] 任锋利[1] 

机构地区：[1]兰州理工大学理学院,甘肃兰州730050

出　　处：《甘肃科学学报》2015年第6期9-13,共5页Journal of Gansu Sciences

摘　　要：研究了响应变量随机缺失下部分线性测量误差(EV)模型的估计问题。利用核估计和最小二乘的方法给出参数β和非参数g的两步估计,并在适当条件下证明了它们的强相合性和渐近正态性,数值模拟表明第二步估计值比第一步估计值更优。The estimation problem in the adaptive variable coefficient EV model with response variables missing is discussed. By using kernel estimation method and least squares method,the parameterβ and non-parameter g are esti- mated by two-step estimation, the strong consistency and the asymptotic normality of the estimated value is proved un- der proper conditions. The numerical simulation shows that the second-step estimate is superior to the first-step one.

关 键 词：部分线性EV模型 缺失数据 强相合性 渐近正态性 

分 类 号：O212.7[理学—概率论与数理统计]