%0 Journal Article
%T Classical and Bayesian estimation of Kumaraswamy distribution based on type II hybrid censored data
%J -
%D 2018
%R DOI Code: 10.1285/i20705948v11n1p235
%X £¿In the literature£¿, £¿different estimation procedures are used for inference about {\color{red} Kumaraswamy} distribution based on complete data sets£¿. £¿But£¿, £¿in many life-testing and reliability studies£¿, £¿a censored sample of data may be available in which failure times of some units are not reported£¿. £¿Unlike the common practice in the literature£¿, £¿this paper considers non-Bayesian and Bayesian estimation of£¿ £¿Kumaraswamy parameters when the data are type II hybrid£¿ £¿censored£¿. £¿The maximum likelihood estimates (MLE) and its asymptotic variance-covariance matrix are obtained£¿. £¿The asymptotic variances and covariances of the MLEs are used to construct approximate confidence£¿ £¿intervals£¿. £¿In addition£¿, £¿by using the parametric bootstrap method£¿, £¿the construction£¿ £¿of confidence intervals for the unknown parameter is discussed£¿. £¿Further£¿, £¿the Bayesian estimation of the parameters under£¿ £¿squared error loss function is discussed£¿. £¿Based on type II hybrid£¿ £¿censored data£¿, £¿the Bayes£¿ £¿estimate of the parameters cannot be obtained explicitly; therefore£¿, £¿an approximation method£¿, £¿namely Tierney and Kadane's approximation£¿, £¿is used to compute the£¿ £¿Bayes estimates of the parameters£¿. £¿Monte Carlo£¿ £¿simulations are performed to compare the performances of the different methods£¿, £¿and one real data set is analyzed for illustrative purposes£¿
%U http://siba-ese.unisalento.it/index.php/ejasa/article/view/18969