Tingkat Kerawanan Desa Berdasakan Dampak Bencana Kab.Bojonegoro Dengan Metode Clustering Algoritma K-Means

Abstract

Kabupaten Bojonegoro merupakan daerah cukup rawan bencana terutama banjir, tanah longsor, dan cuaca ekstrem, tercatat 2.162 kejadian yang menyebabkan kerugian materiil dan moril meningkat signifikan dalam tiga tahun terakhir, mencapai Rp 1.986.210.000. Dalam kaitannya untuk mengurangi kerugian dimasa mendatang dapat dilakukan pencegahan sedini mungkin, dengan dilakukannya penelitian mengenai potensi daerah yang rawan bencana berdasarkan dampak bencana yang terjadi di daerah tersebut. Penelitian ini menggunakan data 430 desa (2019-2022), dengan 6 parameter dampak bencana meliputi jumlah kejadian bencana, korban jiwa, rumah terdampak, lahan terdampak, kerugian harta dan kerusakan fasilitas. Tujuannya mengidentifikasi tingkat kerawanan desa dengan metode Clustering K-Means. Validasi jumlah cluster optimal dilakukan dengan Davies Bouldin Index (DBI) dan data mining mengikuti standar CRISP-DM. Hasil uji coba menunjukkan jumlah cluster optimal k=3. Analisis cluster menunjukkan bahwa cluster 1 (3 desa) mengalami bencana besar lebih banyak merusak rumah dan lahan, cluster 2 (17 desa) mengalami bencana dengan korban signifikan, cluster 3 (410 desa) mengalami bencana dengan dampak paling rendah.The Bojonegoro Regency is a region that is quite prone to disasters, especially floods, landslides, and extreme weather. There have been 2,162 recorded incidents in the past three years, causing significant material and moral losses totaling IDR 1,986,210,000. In an effort to reduce future losses, early prevention measures can be taken by conducting research on the potential disaster-prone areas based on the impact of disasters that have occurred in the region. This study utilized data from 430 villages (2019-2022), considering 6 disaster impact parameters, including the number of disaster events, casualties, affected houses, affected land, material losses, and facility damage. The objective was to identify the vulnerability level of villages using the K-Means Clustering method. The optimal number of clusters was validated using the Davies Bouldin Index (DBI), and the data mining process followed the CRISP-DM standard. The trial results indicated that the optimal number of clusters is k=3. Cluster analysis revealed that Cluster 1 (3 villages) experienced more significant disasters, causing more damage to houses and land; Cluster 2 (17 villages) faced disasters with significant casualties, while Cluster 3 (410 villages) experienced disasters with the lowest impact.