Randomization in survival analysis

In this paper, we estimate the distribution function and the expectation of an increasing process. The data are supposed to be censored. The novel idea is to extend the Kaplan-Meier approach to such processes, by a randomization method. A martingale application is also provided.