This thesis presents a comprehensive exploration of the Fisher Scoring algorithm’s implementation across different link functions within the framework of generalized linear models (GLMs) under the normal distribution assumption. The study includes an evaluation of the performance of various Generalized Linear Models (GLMs) designed with a chosen link function, applicable to insurance data. Data is leveraged from a Swedish motor third-party liability insurance dataset. To check the model’s effectiveness, a comparative analysis is done, comparing the models and their variations, including the exploration of models with interaction within this framework.     «

This thesis presents a comprehensive exploration of the Fisher Scoring algorithm’s implementation across different link functions within the framework of generalized linear models (GLMs) under the normal distribution assumption. The study includes an evaluation of the performance of various Generalized Linear Models (GLMs) designed with a chosen link function, applicable to insurance data. Data is leveraged from a Swedish motor third-party liability insurance dataset. To check the model’s effect...     »