Practical Model for Evaluating the Risk of a Person to Commit a Criminal Offence

As the society develops, the volume of criminal offences and the complexity of crimes also increases. Data mining techniques can be used to help the police departments and explore already existing data to predict for a person the potential risk of committing a criminal offence. This information, known in advance by the authorities, could lead to smart preemptive actions that could lower the crime rates. The purpose of this research is to find a way of modeling the existing information about persons and test what are the machine learning algorithms that give the best results. The work is specific to the Romanian context, the results are very promising, but intensive tests on larger sets of data should be performed.