Analisis Komparasi Clustering Tingkat Penyebaran Demam Berdarah Dengue Dengan Algoritma K-Means Dan K-Medoids

Abstract

Disease is a disorder that occurs in the body both in form and function so that the body cannot work properly or normally. According to WHO (World Health Organization), the number of reported cases of dengue fever increased more than 8 times. In 2000 there were 505,430 cases in the world. Then in 2010, the number of cases jumped to 2.4 million cases. In 2019, 5.2 million cases were reported. Clustering is a method in data mining that aims to group objects into clusters that have similar characteristics to one another. Cluster analysis is an important thing to do to monitor the level of spread of dengue disease so that the government can take action to deal with dengue disease in a focused manner in areas that have a high level of spread. This study utilizes data on the number of cases of dengue hemorrhagic fever and the population in Karawang. The design of the KDD method uses the K-Means and K-Medoids algorithms. Based on the evaluation results, scheme 3, which uses the original dataset or without normalization and uses the K-Means algorithm, is the best scheme with a DBI value of 0.472. Based on the evaluation results, K-Means has a better performance than K-Medoids using either the unnormalized dataset or the normalized dataset.