用于Logistic回归诊断的经验Pearson残差图

目前，回归诊断不仅用于一般线性模型的诊断，还被逐步推广应用于广义线性模型领域（如用于Ｌｏｇｉｓｔｉｃ回归模型），但由于一般线性模型在残差分布的假定等方面有所不同，所以推广和应用还存在许多问题。鉴于此，作提出一种新的图示方法－经验Ｐｅａｒｓｏｎ残差图，用以考察Ｌｏｇｉｓｔｉｃ模型的拟合优度及检测异常点。

Empirical Pearson residual plot in Logistic regression diagnostics

Recently ,regression diagnostics applied not only in general linear models but also in generalized linear models including Logistic regression model. However,there are some problems not overcame due to differences between general linear models and generalized linear models in hypothesized residual distributions and etc. In this paper,a graphical method empirical Pearson residual plot,is proposed for inspectingLogistic regression model tenability and detecting outliers.