Penerapan Association Rule dalam Mengidentifikasi Pola Kriminalitas di Wilayah Hukum Polsek Selebar Kota Bengkulu

Julia Dwi Azizah (Universitas Bengkulu)
Yudi Setiawan (Universitas Bengkulu)
Niska Ramadani (Universitas Bengkulu)

Jurnal Sains dan Teknologi
Vol. 5 No. 4 (2026): Agustus 2026

Crime data stored by law enforcement agencies have primarily been utilized as documentation and incident reports, resulting in the underutilization of information that could reveal relationships among criminal events. This study aims to implement the Association Rule method using the FP-Growth algorithm to identify theft crime patterns and visualize their spatial distribution through a Geographic Information System (GIS). The study utilized theft crime data from the jurisdiction of the Selebar Police Sector, Bengkulu City, with crime location, occurrence time, and type of theft as the primary parameters. The research process included data preprocessing, frequent itemset generation using the FP-Growth algorithm, association rule generation with a minimum support value of 2% and a minimum confidence value of 60%, and visualization of the results using QGIS software. The results indicate that the FP-Growth algorithm successfully identified relationships among crime locations, occurrence times, and types of theft. Several association rules achieved confidence values of up to 100%, while the highest lift value of 2.23 was found at Jl. Pancur Mas–Sukarami. Furthermore, GIS-based visualization successfully illustrated the spatial distribution of crime patterns, facilitating the identification of locations associated with theft incidents.