تقدير معلمات توزيع بواسون المعمم باستخدام أسلوب المحاكاة

Abstract EN

The Poisson distribution has the property of equal dispersion because of the equal variance and mean this characterizes this distribution from among other discrete distributions.

Often and in many real data the population under study are loss of heterogeneous and most of datasets express over dispersion.

In this paper, we introduced the generalized Poisson distribution with two parameters ( ) as one of the characteristics of this distribution is that the population is non-heterogeneous also it provides a model which suits most count data.

The parameters of the generalized Poisson distribution were estimation by different methods, which are Moment method, maximum likelihood method and Mixed (Shrinkage) method.

Also, comparison between those methods through simulation study, with different sample sizes and different values of ( ) using the mean square error (MSE).

The theoretical results are supported by a simulation and lead to that the Mixed (Shrinkage) method is better.