Efficient estimation for the Cox model with varying coefficients

• 成果类型: Article
• 署名作者: Chen, Kani; Lin, Huazhen; Zhou, Yong
• 署名单位: Hong Kong University of Science & Technology; Southwestern University of Finance & Economics - China; Chinese Academy of Sciences
• 刊物名称: BIOMETRIKA
• ISSN/ISSBN: 0006-3444
• DOI: 10.1093/biomet/asr081
• 发表日期: 2012
• 页码: 379392
• 关键词: time-dependent coefficients; local partial-likelihood

摘要

A proportional hazards model with varying coefficients allows one to examine the extent to which covariates interact nonlinearly with an exposure variable. A global partial likelihood method, in contrast with the local partial likelihood method of Fan et al. (2006), is proposed for estimation of varying coefficient functions. The proposed estimators are proved to be consistent and asymptotically normal. Semiparametric efficiency of the estimators is demonstrated in terms of their linear functionals. Evidence in support of the superiority of the method is presented in numerical studies and real examples.