Hurdle strict arcsine model

Abstract

The hurdle model is a finite mixture model where the zeros are generated by a particular distribution while the positive counts are generated by another (truncated) distribution. The discrete distributions commonly considered for hurdle models are the Poisson and negative binomial distributions. The hurdle models are also widely used for over- and under-dispersed count data. In this study, a new hurdle model, which is hurdle strict arcsine model is developed and fitted to two simulated data sets. Maximum likelihood estimation method is used in estimating the parameters.