|
Estimating the parameters of an exponentiated inverted Weibull distribution under type-II censoringKeywords: Exponentiated invertedWeibull distribution , Type II censored sample , Bayesian and non-Bayesian estimations , Symmetric and asymmetric loss functions , Gibbs and Metropolis sampler Abstract: The exponentiated inverted Weibull distribution is a generalization of the exponentiatedinverted exponential distribution as well as the inverted Weibull distribution. In this paper,Bayes and classical estimators have been obtained for two parameters exponentiated invertedWeibull distribution when sample is available from complete and type II censoring scheme.Several Bayesian estimates are obtained against different symmetric and asymmetric lossfunctions such as squared error and LINEX. This was done with respect to the conjugatepriors for two shape parameters. As expected, when the shape parameters are unknown, theclosed-form expressions of the Bayes estimators cannot be obtained. We used an approximationbased on the Lindley and Markov chain Monte Carlo method (MCMC) methods forobtaining Bayes estimates under these loss functions. All Bayesian estimates are comparedwith the corresponding maximum likelihood estimates numerically in terms of their meansquare error values. Finally, real data sets are analyzed for the purpose of illustration.
|