The goal of this thesis is to give an introduction to generalized linear models in non-life insurance. The components of generalized linear models are shown in a general way and then applied to the example of the Poisson claims frequency model. In particular, the maximum likelihood estimator for the Poisson model is derived. It is shown that the maximum likelihood estimator for canonical link functions yields a unique maximum. Three ways to evaluate the goodness-of-fit of a generalized linear model are presented: deviance, various residuals, and the Akaike information criterion. Using a French motor third-party liability dataset, the handling of different types of explanatory variables is shown. Finally, several generalized linear models are built to estimate the frequency of claims, based on the data.     «

The goal of this thesis is to give an introduction to generalized linear models in non-life insurance. The components of generalized linear models are shown in a general way and then applied to the example of the Poisson claims frequency model. In particular, the maximum likelihood estimator for the Poisson model is derived. It is shown that the maximum likelihood estimator for canonical link functions yields a unique maximum. Three ways to evaluate the goodness-of-fit of a generalized linea...     »