摘要
考虑线性回归模型yi=xi′β+iε,i=1,2,…n,其中ε=(1ε,…,nε)′满足E()ε=0,Cov()ε=21σV(V>0),假定{yi}受到另一独立同分布的随机变量序列{iμ}的污染,仅能观察到污染数据,{yi}与{iμ}独立,本文给出污染参数及回归系数的参数估计,并证明在适当的条件下其具有强,弱相合性。
The paper gives a consideration of the regression modelyi=x′iβ+ci,i=1,2…n where ε=(ε1,…,cn)′ with E(ε)=0,σ1^2V(V〉0) and supposes that {yi} be contaminated by another i. i. d. random variable sequence {μi}, and only the contaminated data be observable, where {μi} is independent of { yi}. The estimation of contamination parameter and regression coefficient is obtained. Furthermore, strong and weak consistency is proved under suitable conditions.
出处
《皖西学院学报》
2006年第5期42-44,共3页
Journal of West Anhui University
基金
安徽省省级教学研究项目(JYXM2003159)
关键词
污染系数
污染数据
强相合估计
相合估计
contamination parameter
contamination data
strong consistent estimation
consistent estimation