[Lecture] Donglin Zeng：Efficient Estimation for Semiparametric Regression Models with Interval-Censored Events
Time: 9:00-10:00am, Oct. 29th, 2021, Friday
Venue: ZOOM（ID：83654741182；password：631014）
Spreker: Donglin Zeng, Professor, The University of North Carolina, at Chapel Hill
Abstract:Multivariate interval-censored events are common in practice when disease events of interest contain both symptomatic and asymptomatic events, or the recurrence of the same type event is examined intermittently. Multivariate analysis of interval-censored data makes use of the most information for assessing the effects of risk factors, and is also useful for individual risk prediction. We propose a general modelling framework to analyze such data in which shared random effects are introduced to account for within-subject dependence. Nonparametric maximum likelihood method is adopted for inference, and we devise an efficient algorithm for computation based on latent Poisson processes. The estimators are shown to be consistent, asymptotically normal and semiparametrically efficient. The numerical performance of the proposed method is demonstrated through extensive simulation studies and applications to a cohort study on cardiovascular diseases and a skin cancer trial.
Speaker’s Bio:曾冬林是北卡罗莱纳大学教堂山分校的生物统计系教授。他2001年获密西根大学统计系博士。他的主要研究方向包括半参数模型，个性化治疗，机器学习的统计问题和因果推断。他在各类统计期刊发表近200篇文章。他是数理统计学院和美国统计学会的会士。
