IDENTIFIKASI PENYEBARAN GEMPA DI INDONESIA DENGAN METODE CLUSTERING

Abstract

The spreads of earthquake has been identified by using Clustering methods which are depends on four parameters, i.e. magnitude, depth, location, and elevation. We used EM method, KMeans, and Xmeans as Clustering methods which is using four parameters. The first clustering step based on 4 parameters which used EM method. This method shows the highest earthquake happen at Banda’s sea which is located 127.5747 east longitude and 1.7943 south latitudes. Whereas, KMeans and XMeans method predicts that the highest earthquake located at 95 east longitude and 5 north latitudes in northern Sumatera. The second clustering process was done based on location. This process which is using EM method, KMeans, and XMeans predicted that the most frequent earthquake happened at Sumatera Island which locates at 96 east longitude and 4.5 north latitudes. In the third clustering process, we used two parameters, i.e. magnitude and depth. According to three methods the results show the data’s at earth surface. The last clustering process has been done based on magnitude, depth, and elevation. Generally, by using three methods, the results show the largest earth data’s locate in sea area which has depth about 42 km and the magnitude is 6 scale Richter (SR). These all the clustering processes indicate that EM method is the most appropriate for representing actual the spread of earthquake. It’s because of EM method can be showing the spread of earthquake with the most dense data’s in every cluster based on deviation standard.

Keywords: The spread of earthquake, Clustering methods, EM, KMeans, XMeans.