CERTAIN STATISTICAL PROPERTIES AND APPLICATIONS OF GENERALIZED POISSON- LINDLEY DISTRIBUTION: REVISITED

Abstract

The discrete Poisson-Lindley distribution is a well-known mixture Poisson distribution which has wide applications in the theory of accident proneness investigated by Sankaran (!970) with application to errors and accident data. To increase the flexibility for modelling purpose Mohmoudi and Zakerzadeh (2010) obtained an extended version of the Poisson-Lindley distribution known as Generalized Poisson-Lindley distribution by mixing the Poisson distribution with generalized Lindley distribution. In this paper, an attempt has been made to restudy some statistical properties of the generalized Poisson-Lindley distribution, like recursive formula involving probability function, factorial moments and cumulants etc. The problem of estimation of the parameters is also discussed.Also, the zero modified version and size-biased form of the distribution has been investigated. A set of reported data has been considered for fitting of this distribution and the fit is compared with that obtained from other derived distributions. The results show that, generalized Poisson Lindley distribution using composite method of estimation can provide a better statistical fit than the traditional Poisson Lindley distribution.