|
Activity Number:
|
471
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306190 |
|
Title:
|
A General Semiparametric Transformation Model for Survival Data
|
|
Author(s):
|
Hao Liu*+ and Alexander Tsodikov
|
|
Companies:
|
University of California, Davis and University of California, Davis
|
|
Address:
|
Division of Biostatistics, MS1C, Davis, CA, 95616,
|
|
Keywords:
|
semiparametric model ; transformation model ; proportional hazard model ; proportional odds model ; cure model ; prostate cancer
|
|
Abstract:
|
We study a general class of nonlinear transformation models for semiparametric regression for the survival data. The proposed model for the survival function separates analytically the baseline cumulative hazard function and the regression component. It includes proportional hazard model, proportional odds model, transformation model and cure model as the submodels. We develop the nonparametric maximum likelihood estimation (NPMLE) and prove the strong consistency and asymptotic normality of NPMLE estimators by the theory of empirical processes. The numerical calculation is dealt by the general framework of quasi-EM (QEM) algorithm. We illustrated the methodology by a simulation study and the analysis of a survival data for prostate cancer patients.
|
