自协方差函数的局部影响分析

Local Influence in the Auto Covariance Function

引入广义影响函数及广义Cook统计量,对AR(1)模型的自协方差函数进行扰动分析。运用局部影响分析方法,克服数据删除对时间序列样本数据相依性的破坏和忽略,可以一次性探测出对自协方差函数有较大影响的点。

Based on general influence function and Cook statistic, we investigate the the small perturbation on the auto covariance function. In this paper, local influence analysis which overcome the destruction of dependency and neglect is in- troduced todetect all the influential point on the auto covariance function in AR（1） model at a time.