基于Pena距离的加权最小二乘估计的影响分析  

Influence Analysis on Weighted Least Squares Estimator Based on Pena Distance

作　　者：孟丽丽[1] 卢志义[2] 

机构地区：[1]天津商业大学理学院,天津300134 [2]天津财经大学统计学院,天津300222

出　　处：《数理统计与管理》2009年第2期252-257,共6页Journal of Applied Statistics and Management

摘　　要：利用Pena距离对加权线性最小二乘估计的影响问题进行讨论,得到加权最小二乘估计的Pena距离的表达式,对其性质进行讨论,从而得到高杠异常点的判别方法.文中对Pena距离与Cook距离的性能进行了对比,得到在一定条件下Pena距离优于Cook距离的结论.并通过数值实验对此方法的有效性进行验证.In this paper , the influence analysis of weighted linear least squares estimator is discussed based on the Penn distance. The expression of Penn distance of weighted least squares estimator is obtained and its properties are discussed, meanwhile the discrimination of high-leverage outlier are obtained. The function of the Penn distance and the Cook distance is contrasted , the conclusions are obtained that the Penn distance is better than and the Cook distance on certain conditions. An example is shown to illustrate the validity of the method.

关 键 词：加权最小二乘估计 Pena距离 影响分析 

分 类 号：O212.1[理学—概率论与数理统计]