A bivariate geometric minification integer-valued autoregressive model

Abstract

In this manuscript we introduce a new bivariate minification integer-valued autoregressive model of order one. The model is based on a modification
of the negative binomial thinning operator. The basic features of the model are given, where some of them are used for estimating of the parameters. Two methods are used to estimate the unknown parameters, the conditional maximum likelihood method and the conditional least squares method. The characteristics of the estimates obtained using these two methods are checked through some Monte-Carlo simulations.