Parametric inference of non-informative censored time-to-event data

Abstract

Random or non-informative censoring is when each subject has a censoring time that is statistically independent of their failure times. The classical approach is considered for estimating the Weibull distribution parameters with non-informative censored samples which occur most often in medical and biological study. We have also considered the Bayesian methods via gamma priors with asymmetric (general entropy) loss function and symmetric (squared error) loss function. A simulation study is carried out to assess the performances of the methods using mean squared errors and absolute biases. Two sets of data have been analysed for the purpose of illustration.