摘要
本文对多元线性回归模型定义了AP统计量和距离影响函数;将它们分解为两项之积,指出强影响点、异常点、高杠杆点间的内在联系;讨论了上述三种点的探查方法.最后,给出与距离影响函数有关的一个定理.
In this paper,we define Ap-statistic and distance influence function in multivariate linear regression model, and decompose them into a product of two factors. The decompositions clarify the relationship of influential observations, outliers, and high leverage cases. We also discuss the detective method for these points and give a theorem relative to the distance influence function.
出处
《高校应用数学学报（A辑）》
CSCD
北大核心
1992年第4期495-501,共7页
Applied Mathematics A Journal of Chinese Universities(Ser.A)
关键词
AP统计量
距离影响函数
线性回归
Ap-statistic, Distance Influence Function, Outliers, High leverage Cases, Influential Observations.
