A New Integer-Valued AR(1) Process Based on Power Series Thinning Operator

Abstract

In this paper, we introduce the first-order integer-valued autoregressive (INAR(1)) model, with Poisson-Lindley innovations based on power series thinning operator. Some mathematical features of this process are given and estimating the parameters is discussed by three methods; conditional least squares, Yule-Walker equations and conditional maximum likelihood.Then the results are studied for three special cases of power series operators. Finally, some numerical results are presented with a discussion to the obtained results and Four real data sets are used to show the potentially of the new process.