<![CDATA[This research aims to classify regional police forces in Indonesia based on the level and characteristics of criminal activity by applying the K-Medoids method optimized using the Artificial Bee Colony (ABC) algorithm. This grouping is intended to identify the dominant crime patterns in each region and evaluate the effectiveness of the methods used in producing representative clusters. The analysis results show that the K-Medoids-ABC method produces three main clusters, with the distribution of each consisting of 7 regional police departments in cluster 1, 5 regional police departments in cluster 2, and 21 regional police departments in cluster 3. Cluster validation using the Silhouette Index (SI) yielded a value of 0.387, indicating that the clustering results fall into the weak structure category, meaning the cluster structure is formed but with weak separation (Weak Separation). Cluster 1 shows a moderate and relatively even crime rate, Cluster 2 is dominated by crimes against life and crimes of fraud, embezzlement, and corruption, while Cluster 3 shows low values across all variables, with the lowest values for violent property crimes and drug-related crimes. This cluster reflects regions with relatively safe conditions, as evidenced by very low crime rates. These differences in characteristics between clusters reflect the diversity of factors causing crime in each region and have important implications for formulating more contextual and targeted crime prevention strategies.]]>
Copyrights © 2026
