Generalized Bayesian non-informative prior estimation of Weibull parameter with interval censoring

Interval-censored data consist of adjacent inspection times that surround an unknown failure time. We seek to determine the best estimator for the Weibull scale parameter using interval-censored survival data. Consideration is given to the classical maximum likelihood and Bayesian estimation under squared error loss with interval censoring using noninformative prior and a proposed generalization of non-informative prior. The study is based on simulation and comparisons are made using mean squared error and absolute bias. We find that the proposed generalized non-informative prior is the preferred estimator of the scale parameter.