-----------------------------------------------------------

Statistics Seminar Series

Thursday, September 27, 2007 @ 4:00PM
Cullimore Hall Room 307
New Jersey Institute of Technology

Nonproportional Semiparametric Regression Models

Zhezhen Jin, PhD.

Columbia University

The Cox proportional hazards model (Cox, 1972) is the most widely used semiparametric method for the analysis of censored data. However, the validity of this approach depends on the proportional hazards assumption which can be violated in many important practical occasions. In this talk, I will review recent development in the nonproportional semiparametric models for the analysis of censored data. The models include the accelerated failure time (AFT) model and the linear transformation models. Their extension to the analysis of multiple events data including recurrent events, clustered survival data will be also discussed.