The admissibility of the Kaplan-Meier and other maximum likelihood estimators in the presence of censoring (Q756886)

\textit{G. Meeden}, \textit{M. Ghosh} and \textit{S. Vardeman} [ibid. 13, 811-817 (1985; Zbl 0581.62006)] proved the admissibility of some well-known nonparametric estimators. The arguments proving admissibility involved two essential steps. The first one is based on an argument of \textit{M. P. Cohen} and \textit{L. Kuo} [ibid. 13, 262-271 (1985; Zbl 0575.62011)] by which it is sufficient to prove the admissibility for some multinomial family. The second step then uses the stepwise Bayes technique introduced by \textit{K. Alam} [ibid. 7, 282-283 (1979; Zbl 0401.62010)] to prove admissibility for the multinomial distribution. Here the authors use the same arguments to prove admissibility of the Kaplan-Meier estimator, for the nonparametric estimation of the survival function, when censoring is present. In fact, the admissibility of a whole class of maximum likelihood estimators, which includes the Kaplan- Meier estimator, is demonstrated.