One-inflated Bayesian Mixtures for population size estimation

The phenomenon of one-inflation frequently affects the estimates of population size when the available dare are represented by frequencies of counts. A particular behavioral effect preventing subsequent captures after the first one may be the reason for such an effect. We consider a Bayesian semi-parametric approach by fitting a truncated Dirichlet process mixture model as a base tool for modeling repeated count data and extend this class to include one–inflation. The proposed methodology is briefly illustrated via a real data application.