基于观测方差的加权Stein估计  

Weighted Stein Estimation Based on the Variances of Measurement

作　　者：袁媛[1] 

机构地区：[1]嘉兴学院数学与信息科学学院,浙江嘉兴314001

出　　处：《嘉兴学院学报》2007年第6期41-44,共4页Journal of Jiaxing University

基　　金：嘉兴学院校内青年基金(70106069)

摘　　要：Whittemore(1989)针对在各观测点等误差的情形提出了一种参数估计的方法,其基本思想是:当观测误差来自方差已知的高斯分布时,用带有观测误差的观测样本的Stein估计代替不可观测的协变量的真实值.对于各观测误差的方差不相等的情况,提出了基于观测方差的加权估计方程的概念,并通过数据模拟证明了由加权估计方程解出的估计(记为加权的Stein估计)的优越性.Whittemore （1989） proposed Stein estimation of the unobserved true covariates, provided that the measurement error is Gaussian with known variance, when the variances of the measurement errors are equal in different observed points. The author puts forward the weighted estimating equation based on the variances of the measurement errors when the variances are unequal, the estimation solved from the equation being called the weighted Stein estimation. And its superiority could be proved through the simulation results.

关 键 词：EV模型 STEIN估计 加权估计方程 

分 类 号：O212[理学—概率论与数理统计]