CD Skripsi
Implementasi k-means clustering dan principal components analysis untuk pengelompokan daerah berdasarkan kesiapsiagaan bencana di Sumatra Barat
West Sumatra Province has a high level of disaster vulnerability due to its geographical location in an active seismic zone and tectonic plate convergence area. However, no mapping system is available to classify regions based on disaster preparedness, making disaster management ineffective. Therefore, this research aims to classify and describe the main characteristics of areas in West Sumatra based on disaster preparedness, as a foundation for formulating more specific and targeted disaster management policies. The method used is K-Means Clustering with validation using Silhouette Coefficient, Elbow Method, and Davies-Bouldin Index (DBI), as well as model optimization using Principal Component Analysis (PCA). The data used consists of 19 administrative areas with 20 features related to disaster preparedness. Initial clustering results showed the highest silhouette score of 0.23. PCA was used to optimize the clustering to 0.47. Based on experiments, the optimal number of clusters was found to be k=9 with a DBI value of 0.29. K-Means Clustering successfully identified nine disaster preparedness characteristics of the regions. The implications of this research are expected to provide a scientific foundation for local governments and related agencies in establishing more effective and data-driven disaster management policy priorities, as well as improving community preparedness for disaster risks in West Sumatra.
Keywords: K-Means Clustering, Disaster Preparedness, Principal Component Analysis, Disaster Area Clustering, West Sumatra
Tidak tersedia versi lain