Diseño de perfiles de asegurados en el ramo del automóvil

Abstract

In recent years, insurance companies are increasingly competitive. It is therefore very important that they have a good system of classification of individuals according to their behavior in accident ratio. We will work with empirical data from a Spanish insurance company. It will give an estimate of the role of the risk of occurrence of an incident based on driver age, taking into account other relevant variables. In addition, we get the age ranges that better or worse they behave toward risk. To do this we use the technique of functional principal component analysis and an estimate of the correlation function. With the results obtained by this functional technique we will conduct a cluster analysis to identify homogeneous groups in the behavior of claims. This would allow insurance companies to establish a tariff for each of the groups.