Comparative study of the performance of a non-parametric self-consistent estimator with that of maximum likelihood estimator under random censorship of a bivariate survival function

Abstract

Traditional approaches are dealt with for data analysis in bioreliability based on parametric or non-parametric statistical techniques. Parametric techniques for censored data lead to cumbersome results and too many complications; the non-parametric techniques also (which used to be employed as an alternative to parametric procedures), according to Leurgans et al. (1984), do not converge in probability to the true values of the parametric. Therefore, it was necessary to re-examine the issue, namely how a comparison between parametric and non-parametric techniques (with known values of the parameter) stands. The investigation is based on the same approach.