A Bayesian bootstrap for censored data (Q2366729)

In \textit{D. B. Rubin}'s [ibid. 9, 130-134 (1981)] Bayesian bootstrap of an empirical distribution function the original observations obtain new (random) masses from the vector of spacings associated with \(n-1\) independent uniform random variables. The present paper extends Rubin's Bayesian bootstrap to the Kaplan-Meier estimator in a random censorship model. It is shown that the resulting process has the same limit behavior as the original Kaplan-Meier process.