Stanovení pojistného u havarijního pojištění pomocí regresních modelů

Abstract

A model for insurance premium pricing in non-life insurance sphere, concretely collision insurance, establishing with regression models is an object of this thesis. Text consist of theoretical and practical part. Theoretical part consist of two main chapters. The first chapter is focused on clearing insurance and related terms up and models in insurance and insurance premium calculation proces defining. Thesis is primarily focused on non-life insurance, which is main principle of it. In the second chapter the regression models with an emphasis on the models of binary options and proces of regression models establishing are defined. In the practical part a knowledge from the theoretical part is aplicated. First of all are individual variables analysed, modified and tested with regards on corect specification and significance. Than a regression model is estimated and tested with regards on corect specification and significance, and the risk conected with every insurance related is quantified, thus a probability of the insured event occurrence is quantified. Than the model is evaluated with regards on suitability for insurance premium pricing and success in pricing. At the end the annual netto insurance premium is quantified for each tarif cell, and the final model is evaluated again. The result of this thesis is individual model for natural persons and legal entities.