Maximum likelihood estimation of a survival function under the Koziol- Green proportional hazards model (Q1094064)

The authors study the maximum likelihood estimator (MLE) of the survival function of the lifetime of interest in a proportional hazards competing risks model. Large sample properties such as uniform mean square consistency, the LIL and weak convergence to a Gaussian process are derived. The MLE is also shown to be asymptotically more efficient than the Kaplan-Meier estimator under the specified model.