The McDonald Weibull model
Gauss M. Cordeiro, Elizabeth M. Hashimoto, Edwin M. M. Ortega
AbstractFor the first time, we propose a five-parameter lifetime model called the McDonald Weibull distribution to extend the Weibull, exponentiated Weibull, beta Weibull and Kumaraswamy Weibull distributions, among several other models. We obtain explicit expressions for the ordinary moments, quantile and generating functions, mean deviations and moments of the order statistics. We use the method of maximum likelihood to fit the new distribution and determine the observed information matrix. We define the log-McDonald Weibull regression model for censored data. The potentiality of the new model is illustrated by means of two real data sets.Keywords: Kumaraswamy distributionmaximum-likelihood estimationMcDonald distributionmean deviationmoment generating functionquantile functionWeibull distribution AcknowledgementsThis work was supported by FAPESP grant 2010/04496-2 and CNPq, Brazil.…