Speakers
Parallel Workshops
Recidivism is a major issue for society and law enforcement, with several factors contributing to the likelihood of repeat offending. Offender profiles can be analyzed using existing prison databases to identify potential recidivists. Artificial intelligence can aid in understanding underlying patterns and critical factors that contribute to recidivism. In this study, machine learning algorithms were employed to predict recidivism using three datasets: US national correction data, NIJ challenge data from Georgia prisons, and recidivism data from the government of Catalonia. The results showed an average accuracy of 75% for the US national correction data, 69% for the Georgia prison data, and 66% for the Catalonia juvenile data.