In contemporary society, Machine Learning, which is the field of Artificial Intelligence that exploits statistical algorithms for acquiring from previously collected data in order to find general rules and, thus, perform predictions, is becoming increasingly more present in everyday life. In particular, it has found major employment in various fields such as Computer Science and Finance. Therefore, the application of Machine Learning techniques represents a powerful tool which can be used to expedite a large number of tasks. This study describes and analyses the creation of an ML model intended to forecast claims propensity for fraud in the context of credit insurance.