期刊
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
卷 96, 期 455, 页码 1022-1030出版社
AMER STATISTICAL ASSOC
DOI: 10.1198/016214501753209004
关键词
binomial regression; influence function; M-estimators; model selections; Poisson regression; quasi-likehood; robust deviance; robustness of efficiency; robustness of validity
By starting from a natural class of robust estimators for generalized linear models based on the notion of qua-si-likelihood, we define robust deviances that can be used for stepwise model selection as in the classical framework. Wc derive the asymptotic distribution of tests based on robust deviances, and we investigate the stability of their asymptotic level under contamination. The binomial and Poisson models are treated in detail. Two applications to real data and a sensitivity analysis show that the inference obtained by means of the new techniques is more reliable than that obtained by classical estimation and testing procedures.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据