Analisis Klaster Provinsi di Indonesia Berdasarkan Indikator Mitigasi Bencana dan Karakteristik Wilayah Menggunakan Metode Fuzzy C-Means

Abstract

The high frequency of natural disasters across Indonesia underscores the urgent need for a structured, data-driven mitigation system. Situated at the convergence of active tectonic plates, each region in Indonesia exhibits varying levels of disaster preparedness and response capacity. This study applies the Fuzzy C-Means method to cluster Indonesia’s 38 provinces based on their natural disaster mitigation capacity. Variables used include tsunami early warning systems, evacuation routes, number of disasters, number of sub-districts, land area, and population, all sourced from the Central Statistics Agency in 2024. Prior to clustering, data underwent correlation analysis and normalization through the Min-Max Scaler method. Two clusters emerged: the first representing provinces with high mitigation capacity comprising 5 provinces (13.16%), and the second representing those with low mitigation capacity comprising 33 provinces (86.84%). Cluster quality was measured using the silhouette coefficient, yielding a score of 0.744, which reflects a very strong cluster structure. These findings are intended to support the government in formulating more targeted disaster mitigation policy priorities.