Clustering Analysis of Local Earthquake in Toba Geopark based on K-means and K-Medoids Algorithm

This study investigates the application of K-means and K-Medoids clustering algorithms to local earthquake data from Geopark Toba, a region prone to significant seismic activity due to its tectonic setting. By analyzing seismic events from June 22 to June 24, 2024, the research aims to identify patterns and enhance disaster preparedness. The algorithm provided distinct advantages of K-Means exhibited identical Silhouette Scores (0.6613), Davies-Bouldin Index (0.4893), and Calinski-Harabasz Index (89.5896), indicating that both methods generated similarly well-separated and compact clusters. K-Medoids, on the other hand, showed a lower Silhouette Score (0.4527), reflecting less compact clusters, but outperformed in terms of the Calinski-Harabasz Index (100.7853), which suggests better cluster separation and overall quality. Evaluation using the Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index indicated that K-Means performed the best, capturing dense regions and effectively managing noise The findings contribute to the broader field of seismology, offering insights into earthquake clustering that could inform future research and improve seismic hazard assessments in the region.