IMPLEMENTASI EUCLIDEAN DAN CHEBYSHEV DISTANCE PADA K-MEDOIDS CLUSTERING
Abstract
Clustering is one of the data mining techniques for grouping a set of data with similar characteristics in a cluster. Methods clustering, one of which is K-Medoids or PAM (Partitioning Around Medoids), an algorithm identical to k-means because it breaks the data set into groups. The algorithm k-medoids use to overcome the weakness of k-means which is sensitive to noise and outliers. The level of similarity of characteristics in clustering determines by measuring the distance between the data. This study uses Euclidean Distance and Chebyshev Distance to determine the cluster from each distance measure. This study compares Euclidean and Chebyshev distance implementation on k-medoids clustering. In this study, the group obtained using the Chebyshev distance on k-medoids produces a more optimal cluster.