PENERAPAN K-MEANS CLUSTERING UNTUK SEGMENTASI WILAYAH BERDASARKAN JUMLAH KEJADIAN BENCANA ALAM DI SUMATERA UTARA

North Sumatra is a disaster-prone province in Indonesia, frequently affected by floods, landslides, and extreme weather events. This study employs the K-Means Clustering algorithm to classify 33 districts/cities in North Sumatra based on disaster frequency. The results reveal three distinct clusters: Cluster 0 (low risk), Cluster 1 (high flood and extreme disaster risk), and Cluster 2 (high landslide risk). Model validation using a silhouette coefficient of 0.63 and the elbow method confirms the reliability of the clustering. These findings provide a scientific basis for region-specific disaster mitigation strategies, emphasizing infrastructure development for flood-prone areas and slope stabilization for landslide-prone zones. Future research should incorporate socioeconomic factors and temporal analysis to enhance disaster risk assessment.