Modification of classical clustering algorithms

Abstract

In this paper, a modification of classical clustering algorithms is presented. In this paper, I present a method by which clustering algorithms themselves determine cluster boundaries, the number of groups in which to break down the elements of a data set. Clustering is a data mining method where similar elements are placed in the same cluster, while different elements are placed in a separate cluster. In this paper, a partition algorithm (K-Means) and hierarchical methods (Single Linkage, Complete Linkage, Average Linkage, Ward, Centroid) are presented. The running results show that the clustering algorithms were more or less able to form the clusters without waiting for the cluster number as input.