Estimation of a population distribution based on classification probability model: application in the case of hunting gun crimes in Greece

Abstract

This work has a goal to suggest two iteration processes and the sufficient conditions under which
these processes converge and have useful properties. In other words it investigates the probability and the
sufficient conditions that have to be satisfied. The use of maximum likelihood equation as well as of the
iteration processes that are usually applied is justified only in large samples. An original sample of
hunting gun crimes in Greece led the author, based on a discrete probability model, to estimate the
distribution of five categories of crimes. Some comparisons are made between the suggested methods
concerning the speed of their convergence and accuracy of their results.