PENERAPAN DATA MINING MENGGUNAKAN ALGORITMA K-MEANS UNTUK KLASIFIKASI PENYAKIT ISPA

Abstract

With the rapid development of technology in all fields, from the government, education, agriculture and especially in the health sector, technology can provide fast and accurate information for health teams, doctors, nurses and even patients themselves to make it easier to control their own health. Acute Respiratory Infections (ARIs) are diseases of the upper or lower respiratory tract, usually contagious, that can cause a wide spectrum of diseases ranging from asymptomatic illness or mild infections to severe and deadly illness, depending on the causative pathogen, environmental factors and factors host. The purpose of this study was to apply the K-Means method to classify ARI diseases and to obtain accurate and fast accuracy in classifying symptoms of ARI using the K-Means method. The method used is data mining techniques using the K-Means algorithm. This process resulted in 3 clusters, namely cluster C1 (Regular ISPA) with 81 members, cluster C2 (moderate ISPA) with 103 members, and cluster C3 (Heavy ISPA) with 66 members. It can be seen that the largest number of ARI patients are patients with mild ARI symptoms. Based on the results of the percentage analysis for each cluster, the first cluster has a percentage of 35% of data, the second cluster is 45% of data and the third cluster is 20% of data. Testing using the DBI (Davies Bouldin Index) validation obtained values for each cluster. Testing cluster 1 produces DBI value -0.244, cluster 2 DBI value -0.250, cluster 3 DBI value -0.239. Because the DBI value of cluster 3 is smaller, the cluster can be called optimal.