Poisson回归模型误差方差的估计  

Estimation of the Error Variance of the Poisson Regression Model

作　　者：舒常江 SHU Chang-jiang(Mathematics Teaching and Research Department,Guilin Institute of Information Technology,Guilin 541004,China)

机构地区：[1]桂林信息科技学院数学教研部,广西桂林541004

出　　处：《南宁师范大学学报（自然科学版）》2022年第4期33-36,共4页Journal of Nanning Normal University：Natural Science Edition

基　　金：桂林电子科技大学信息科技学院2020年科研项目(XJ202025)。

摘　　要：Poisson回归模型是一类常见的广义线性回归模型,它在数据处理中有广泛的应用,且其误差方差在模型诊断等方面有重要作用.该文利用残差平方和对随机情形的Poisson回归模型的误差方差进行估计,证明了在该方法下模型的误差方差具有渐近正态性,并求出了传统残差平方和的渐近方差.Poisson regression model is a common class of generalized linear regression models. This model is widely used in data processing, and its error variance plays a significant role in model diagnosis. In this paper we estimate the error variance of Poisson regression model in random cases by the sum of squares of residuals, prove the asymptotic normality of the error variance by the same method, and obtain the asymptotic variance of the sum of squares of traditional residuals.

关 键 词：Poisson回归模型 广义线性模型 误差方差 残差平方和 

分 类 号：O212.1[理学—概率论与数理统计]