AN INTELLIGENT MODEL FOR DETECTING ASSOCIATION RULES IN CRIMINAL DATABASES

Abstract

This study explores patterns in criminal activities through association rule mining analysis of criminal offenses in Ukraine's Ternopil region from 2012 to 2023. Using the FP-Growth algorithm and RapidMiner Studio, the research reveals significant patterns in criminal behavior by examining various factors including crime type, lighting conditions, and group dynamics. The analysis uncovered several key patterns: groups frequently commit theft regardless of lighting conditions, with Ternopil as a primary location for these crimes. Group criminal activities show peak occurrence in April and during the latter part of the week, particularly on Thursdays and Fridays. A notable finding is the close association between theft and illegal vehicle appropriation, with vehicle theft commonly executed by groups under dark conditions.

This study's comprehensive approach incorporates multiple factors to provide a more complete understanding of criminal activity patterns. The discovered association rules offer valuable insights for law enforcement agencies, enabling them to develop targeted prevention strategies and optimize resource allocation. The findings particularly highlight the connection between theft and illegal vehicle appropriation, with the latter typically occurring during nighttime hours, suggesting the need for increased patrol presence in streets and parking areas during these times. The study acknowledges that crime patterns are often specific to particular localities, indicating the need for additional research across different regions to develop a more comprehensive understanding of criminal behavior and its influencing factors. The research demonstrates that data mining techniques, specifically association rule mining, can effectively uncover non-obvious yet meaningful patterns in crime data. These insights can inform proactive policing efforts, ultimately contributing to enhanced crime prevention and safer communities. As data-driven approaches continue to evolve, their integration into law enforcement practices will become increasingly crucial in the effective combat against criminal activities.