Applying K-Means Algorithm for Clustering Analysis Earthquakes Data in West Nusa Tenggara Province

Abstract

This study aims to cluster and visualize the earthquake data on a geographical map to determine earthquakes' characteristics using the k-means algorithm. Cluster analysis using the k-means algorithm was carried out on the earthquake data. K-means algorithm is familiar and is one of the well-known techniques to have been applied in cluster analysis. One of Its advantages in cluster analysis is scaling large datasets, for example, earthquake data. The data used in this study is earthquake data in the West Nusa Tenggara from 1991 to 2021. Applying the proposed k-means algorithm, the optimal number of clusters (k) used in this clustering is 2, based on the highest silhouette score of 0.749. The cluster analysis on the geographical map showed that the epicenters of the earthquakes were pretty spread out before 2018, and the number of earthquakes in the eastern region of West Nusa Tenggara is more than in the western area. However, in 2018, the clusters were all bunched in the northern Lombok region. There were a few earthquakes in the west region in 2018, but they happened before August 5. Even after 2019, most earthquakes continue to occur, with the epicenters clustered close to the northern Lombok region