Klasifikasi Kedalaman Kejadian Gempa Menggunakan Algoritma KMeans Clustering: Studi Kasus Kejadian Gempa di Sulawesi

Abstract

Sulawesi is a region that has complex geologic conditions, so disasters caused by the high magnitude of the earthquake occur at any depth. The depth and magnitude of the earthquake causing the disasters had investigated by K-Means Clustering of the machine learning technique. Longitude, latitude, magnitude, and depth attributes used to klaster the earthquake events in the 1970-2022 periods. The klaster number optimized by the Elbow method had validated by the Davies-Bouldin index (DBI). Based on the results, the optimal number of klaster is three klaster, and its Davies-Bouldin index is 0.397. The depth of the first klaster is less than equal to 120 km (shallow earthquake), the second klaster is between 120 km and 350 km (intermediate earthquake), and the third klaster is more than 350 km (deep earthquake). The klaster visualizations of the earthquakes revealed that shallow earthquakes with above 6 SR is frequently occurrence in shallow depth. These is revealed that some Sulawesi Province in the first klaster is vulnerable to earthquake hazard, and the K-Means clustering algorithm is successfully klasters earthquake depth.