Clustering Flood Prone Areas in Deli Serdang Regency Using Density-Based Spatial Clustering Of Application With Noise (DBSCAN) Method

Deli Serdang Regency is the most frequently flooded area in North Sumatra Province, causing many casualties and other losses to residents in flooded areas. Deli Serdang Regency has 22 sub-districts, each of which has a different level of flood vulnerability. Efforts are needed to categorize the level of flood vulnerability that needs to be watched out for in Deli Serdang Regency. The clustering used in this research is Density-Based Spatial Clustering Applications with Noise (DBSCAN). The purpose of this study is to determine the level of proneness to flooding in each region in 2022 in Deli Serdang Regency. The clustering results in this study concluded that using the DBSCAN algorithm, we obtained 2 clusters and 4 noise with a silhouette coefficient value of 0.395050089 with Epsilon 1.19 and MinPts 3. From the silhouette coefficient results, it can be concluded that the cluster structure obtained is weak because, with more variables, the calculation of distance based on density becomes invalid.