<![CDATA[Floods rank among the most common natural disasters occurring in Indonesia, especially Kalimantan Island, where all regencies/cities experienced floods in 2025. Floods are influenced by complex interactions between natural and human factors and require robust clustering methods such as DBSCAN and HDBSCAN to handle noise. In density-based clustering, core points are dense observations satisfying the minimum neighborhood requirement, boundary points are observations near cluster edges, and noise points are isolated observations not assigned to any cluster. The objective of this study is to compare the performance of DBSCAN and HDBSCAN for clustering flood-prone areas on the island of Kalimantan based on DBCV values. The method used in this study is parameter optimization using DBCV between the DBSCAN and HDBSCAN algorithms on 2025 Village Potential (PODES) data from BPS, covering 56 regencies/cities across five provinces on the island of Kalimantan. Six variables were analyzed: flood, water pollution, river/drainage maintenance, forest/land fires, deforestation, and trash disposal. DBSCAN was tested with epsilon of 33.1, 33.2, 33.3, 33.4, 33.5 and MinPts of 2, 3, 4, while HDBSCAN was tested with MinPts of 2, 3, 4. The results of this study indicate that DBSCAN with epsilon of 33.3 and MinPts of 2 produced 3 clusters and 12 noise points, achieving a superior DBCV value of 0.276. Meanwhile, HDBSCAN with MinPts set to 2 produced 15 clusters and 10 noise points, achieving a DBCV of 0.261. Thus, DBSCAN is the best algorithm for clustering flood-prone areas on the island of Kalimantan. The results of this research demonstrate the effectiveness of DBSCAN for flood-prone area delineation, informing targeted mitigation strategies across Kalimantan.]]>
