Στοχαστικά μοντέλα για τον αριθμό των απαιτήσεων: Εφαρμογή στον κλάδο αυτοκινήτου.
Stochastic models for claim claims: Application in automobile insurance
Μαρούλη, Αναστασία Σ.
By stochastic system we mean any real system, the function of which is greatly influenced by luck. In other words, each model is stochastic, the future behavior of which cannot be predicted precisely but only probabilistic. These models are a simulation of the real system, where significant relationships between elements are replaced by corresponding mathematical, while non- significant are ignored. Stochastic processes are adequate mathematical models for the description and study of stochastic systems. This thesis is a study of specific models with regard to their adaptation to data portfolios automobile insurance industry. In actuarial science, many different distributions have been suggested for the description of the number of claims of an insurance portfolio that can be applied instead of the Poisson distribution. In this family of distributions belong the Negative-Binomial distribution, the Poisson-Inverse Gaussian distribution, the strict-arcsine distribution, the Negative Binomial- Pareto distribution, the Poisson–Log Normal distribution and many others. The aim of this dissertation is to write down and study the basic characteristics of the aforementioned distributions and the evaluation of their fitness upon data from the automobile insurance industry. Πανεπιστήμιο