Artificial Intelligence against Serial Offenders

Aleksey Bessonov
DOI: https://doi.org/10.18254/s207751800027535-7
2023-01-01
Artificial Societies
Abstract:The article deals with the use of artificial intelligence algorithms in the investigation of serial crimes committed for sexual reasons, including murders. Based on a database of 1100 serial crimes (Chikatilo, Popkov, Makarenkov, etc.), the author has built their digital model containing information about their typical features and regular relationships between them. Based on this digital model, using artificial intelligence methods, it is possible to build a search portrait of a serial offender, establish a connection between unsolved crimes for seriality and identify the involvement of persons previously prosecuted for similar acts. In working with the data, a number of artificial intelligence algorithms were used: logistic regression, decision trees, random forest, gradient boosting, clustering, neural networks. The results of the study allowed us to draw conclusions about the possibility of using artificial intelligence in the study of crimes, the formation of their digital model, the construction of decision support systems for scientific and practical problems.
What problem does this paper attempt to address?