EVALUASI HASIL KLASTER PADA DATASET IRIS, SOYBEAN-SMALL, WINE MENGGUNAKAN ALGORITMA FUZZY C-MEANS DAN K-MEANS++

Abstract

One technique in data mining is clustering. There are two types of clustering algorithms, namely soft and hard clustering clustering. K-Means++ is hard clustering algorithm which is an improvement of the algorithm K-Means. In K-Means++, center point selection is determined by the concept of probability do not choose at random like at the algorithm K-Means. Algorithm Fuzzy C-Means (FCM) is one of the improvements of the algorithm K-Means. FCM is soft clustering algorithm which applies fuzzy approach to determine the clusters based on the degree of membership. In this study will be evaluated on a cluster results of FCM algorithm and K-Means ++ at the dataset Iris, Wine and Soybean-Small. Results cluster of both algorithms will be compared and will look for the best. The test results of the cluster using the Confusion Matrix and Silhouette Coefficient. The result shows that the algorithm FCM and K-Means have almost similar performance. At the dataset Soybean-Small, Wine both algorithms have the same Silhoutte Coefficient, K-Means++ algorithm has an accuracy rate superior to the FCM algorithm