Estimation of survival curves from dependent censorship models via a generalized self-consistent property with nonparametric Bayesian estimation application (Q1078944)

The author generalizes the concept of self-consistency [cf. \textit{B. Efron}, The two-sample problem with censored data. Proc. 5th Berkeley Symp. math. Stat. Probab., Univ. Calif., 1965-1966, Vol. IV, 831-853 (1967)]. This generalization is then used in order to obtain a family of estimators of the survival function based on right-censored observations, which admit the possibility that the censoring variables are correlated with the true failure times. By assuming a Dirichlet process as a prior distribution, nonparametric Bayes estimators are derived for a specific loss function. Large sample properties of this family of nonparametric Bayes estimators are derived.