Bayesian Estimation & Prediction Based on Complete & Censored Samples

In this book efforts have been devoted in finding a suitable objective prior for the parameters, by using an informative prior density on the unreliability level at a prefixed time, where the analyst does not think in terms of the parameters, but express his technical knowledge in terms of more practical concepts such as the unreliability level at a given time. Then he must convert his information into a degree of belief about the distribution parameters. Thus, in this book two Bayes procedures are applied, when the underline lifetime distribution has the type-II extreme value distribution (EVD); the first procedure used the non-informative prior (NIP) for the purpose of comparison, when there is no information available about the parameters other than that contained in the empirical data. The second procedure, which is the proposed one, used the informative prior (IP), when there is prior information on practical quantities for which an engineer can express his prior knowledge in terms of more practical concepts such as the unreliability level F(t), at a prefixed time.