变系数回归模型的局部M-估计  

Local M-Estimation for Varying-coefficient Linear Regression Model

作　　者：卢静[1] 张日权[1] 

机构地区：[1]太原理工大学理学院,山西太原030024

出　　处：《太原理工大学学报》2008年第2期213-216,共4页Journal of Taiyuan University of Technology

摘　　要：讨论了变系数回归模型的系数函数的估计问题。在通常的局部M-回归方法基础上嵌入一个变窗宽对模型的系数函数进行了估计,并在样本独立同分布的情况下,讨论了系数函数估计的弱相合性和渐近正态性,最后,给出了估计的渐近性质的证明。In this paper, we study the estimation for coefficient functions of the varying-coefficient linear regression model. We apply the local M-estimation augmented with variable bandwidth and get estimation for coefficient functions. The weak consistency and asymptotic normality of estimation on the coefficient functions are given under the condition which the sample is independent identically distributed. At last, proofs are given about the asymptotic properties of the proposed estimation.

关 键 词：变系数回归模型 局部M-估计 稳健性 变窗宽 渐近正态性 

分 类 号：O212.7[理学—概率论与数理统计]