Back To Index Previous Article Next Article Full Text Supplement


Statistica Sinica 20 (2010), 1025-1041





ANALYSIS OF MULTIVARIATE FAILURE TIME DATA

USING MARGINAL PROPORTIONAL HAZARDS MODEL


Ying Chen, Kani Chen and Zhiliang Ying


Shanghai University of Finance and Economics,
Hong Kong University of Science and Technology and Columbia University


Abstract: The marginal proportional hazards model is an important tool in the analysis of multivariate failure time data in the presence of censoring. We propose a method of estimation via the linear combinations of martingale residuals. The estimation and inference procedures are easy to implement numerically. The estimation is generally more accurate than the existing pseudo-likelihood approach: the size of efficiency gain can be considerable in some cases, and the maximum relative efficiency in theory is infinite. Consistency and asymptotic normality are established. Empirical evidence in support of the theoretical claims is shown in simulation studies.



Key words and phrases: Alternating projection, counting process martingale, marginal likelihood, martingale residual, semiparametric efficiency.

Back To Index Previous Article Next Article Full Text Supplement