Analisis Pengelompokan Jenis Kejahatan di Sumatera Utara Berdasarkan Pola Kejadian Tahunan Menggunakan Algoritma K-Means Clustering

This study aims to cluster various types of crimes occurring in North Sumatra Province based on annual incident patterns using the K-Means clustering algorithm. The data utilized are secondary data obtained from the Central Bureau of Statistics (BPS), comprising 34 types of crimes recorded from 2007 to 2021. Prior to clustering, data were normalized using the Z-score standardization method to ensure uniform scaling across variables. The optimal number of clusters was determined using the Elbow Method and Silhouette Plot. The analysis results indicate that four clusters (k = 4) provide the best balance between model complexity and clustering quality. Each cluster reveals distinct crime patterns in terms of frequency and trend stability over the years. The clustering results offer a clearer understanding of crime characteristics in the region and can serve as a foundation for more targeted policy-making, such as resource allocation for law enforcement and data-driven crime prevention strategies. This study demonstrates that data mining approaches, particularly the K-Means algorithm, can significantly contribute to a systematic and comprehensive understanding of crime patterns.