Inference with partially missing causes competing risks data under unit inverse Weibull distribution
摘要
We make inferences considering the competing risks model under the assumption that lifetimes follow unit inverse Weibull distribution with individual failure causes being independent and partially known. Some useful properties of maximum likelihood estimators of parameters are proved under generalized progressive hybrid censoring. Bayesian estimates are derived by using a general class of prior distributions. For the estimation purpose, the cases of both unrestricted and order-restricted parameters are considered. Comparative studies of proposed point and interval estimators are presented using extensive simulation studies. Further, keeping illustration of proposed methods in priority, a useful real life example based on electrode failure times is discussed. At the end, some concluding remarks are made with regard to the considered model.