An analysis of natural disaster data by using K-means and K-medoids algorithm of data mining techniques

Abstract

Indonesia is one of the countries with diverse morphology of the lands, high mountains, and the tropical climates of frequent high rainfall. This condition often causes natural disasters in some areas of the country, which sometimes are so terrible that make a lot of people are missing and suffering. In order to reduce the impact of natural disasters to the people and environment, a research was conducted by capturing data showing the occurrence of the disasters and data about the weather conditions for the last five years. Data is obtained from the official sites of Indonesian National Board for Disaster Management (BNPB) and Indonesian Agency for Meteorological, Climatological, and Geophysics (BMKG). This data is then analyzed by using clustering data mining techniques i.e. k-means algorithm and k-medoids algorithm. The two methods are frequently used to make some analysis of data to find some hidden information. The result shows that weather is not the only factor causing natural disaster. By using the result, the government can make some plans for natural disaster mitigations.