Bayesian population size estimation by repeated identifications of units. A semi-parametric mixture model approach

The use of mixture models for estimating the size of an elusive population when capture rates vary among individuals has received strong attention from researchers involved in multiple system estimation. In this paper we propose a Bayesian semi-parametric approach by considering a truncated infinite dimensional Poisson mixture model for capture recapture count data. An application in official statistics regarding the estimate of the size of criminal populations is used to illus- trate the proposed methodology.