Clustering Of Natural Disaster-Prone Areas In East Java Province Using Fuzzy C-Means Method

Abstract

Natural disasters are one of the most common problems in Indonesia. According to Data Informasi Bencana Indonesia (DIBI) from 2020 to 2024, East Java Province ranks third in terms of the frequency of natural disasters on the island of Java. Systematic mapping of disaster-prone areas is essential to support more effective mitigation and management efforts. This study aims to group 38 districts/cities in East Java Province based on their vulnerability to natural disasters. The methods used are Fuzzy C-Means for the grouping process and Silhouette Coefficient as a tool for evaluating cluster quality. The data used is secondary data with indicators of the number of incidents, number of victims, and amount of damage caused by floods, landslides, extreme weather, drought, earthquakes, volcanic eruptions, and forest and land fires. The clustering results produced three clusters, namely areas with high, medium, and low vulnerability levels. The clustering results were evaluated using the Silhouette Coefficient with a value of 0.2807, which indicates that the clustering results are in the fairly good category and indicate limitations in cluster separation due to the use of limited indicators and overlapping characteristics between regions. Nevertheless, the results of this study can contribute as a basis for consideration in the formulation of disaster mitigation policies, especially in determining intervention priorities and strengthening preparedness in areas with high disaster vulnerability.