BETA TRANSFORMATION OF THE LINDLEY POISSON DISTRIBUTION FOR OVER-DISPERSED COUNT DATA

Abstract

A new distribution for over-dispersed count data is proposed, and its properties are studied. This is a two-parameter distribution which is obtained by introducing an additional parameter beta into the Poisson-Lindley distribution. The goodness-of-fit of this distribution is compared with other distributions that have been proposed to model overdispersion. Two illustrative examples are presented to show the flexibility of the model.