Association Rules Mining in Crime Data Analysis
Olha Ya. Kovalchuk, Serhiy Banakh, Mariia Masonkova, Andrii Kolesnikov, Pavlo Chopyk, Pavlo Basistyi · 2024
This study investigates patterns and trends in criminal activities by employing association rule mining on a comprehensive dataset of 1,975 criminal cases from the Ternopil region of Ukraine between 2012 and 2023. The analysis focused on offenses such as theft, robbery, and illegal vehicle appropriation, considering factors like location, time, lighting conditions, and group involvement. Utilizing the FP-Growth algorithm and RapidMiner Studio, a set of significant association rules was generated, revealing insightful relationships among the variables. The key findings indicate that theft is frequently committed by groups irrespective of lighting conditions, with Ternopil being a hotspot for such crimes. Group criminal activities tend to peak in April and during the latter half of the week, especially on Thursdays and Fridays. Notably, theft and illegal vehicle appropriation were found to be closely associated offenses, with vehicle theft often perpetrated by groups under dark conditions.