For an insurance company, the pricing of its products is an essential part of the business. A useful statistical aid for that process are generalized linear models. They are a generalization of the classical linear model, which corresponds to a linear regression, which has some helpful properties. This thesis deals with the construction of a generalized linear model, the distributions of the exponential family, which are also relevant for insurance models, as well as the key figures of the adaptation quality of the models. Using a statistical tool, a database is processed and a tariff structure is created. In this process a risk premium net is calculated with additional factors to calculate the individual premium for a policyholder. Since data contain a certain uncertainty and error rate, it is recommended to generate an additional risk premium. This risk premium can be calculated with the gaussian error propagation.