重庆理工大学学报（自然科学版）2017，Vol.31Issue(8)：174-181,8.DOI:10.3969/j.issn.1674-8425(z).2017.08.029
右删失数据下广义线性模型的统计诊断
Statistical Diagnostics for Generalized Linear Models with Right-Censored Data
摘要
Abstract
This paper investigates statistical diagnosis problem of the generalized linear models with right-censored data.First, we derive the likelihood function under right-censored data to obtain maximum likelihood estimates for the parameters.Based on the case-deletion models and using the first order Taylor approximation of parameter estimates, we then propose the diagnostic tools such as the generalized cook distance and the likelihood distance to determine outfielders and/or influential cases in the data.We also prove the equivalence of two distances.Finally, we use a real data example to verify the efficiency and feasibility of the proposed diagnostic methods.关键词
广义线性模型/右删失数据/极大似然估计/广义Cook距离/似然距离Key words
generalized linear model/right-censored data/maximum likelihood estimation/generalized cook distance/likelihood distance分类
数理科学引用本文复制引用
季文奇,冯予..右删失数据下广义线性模型的统计诊断[J].重庆理工大学学报（自然科学版）,2017,31(8):174-181,8.基金项目
国家自然科学基金资助项目(11271189) （11271189）
