Optimasi Kinerja Algoritma K-Means pada Penentuan Data Centroid dengan Menggunakan Algoritma Agglomerative Hierarchical Cluster

Abstract

The initial process of clustering the K-Means algorithm is to determine the initial cluster center point (centroid). The selection of the initial centroid in the K-Means algorithm greatly determines the output of the clustering process. The selection of centroids that is often done in the K-Means algorithm is random. However, in this study, the selection of centroids in the K-Means algorithm was carried out by taking the highest data from the AHC (Agglomerative Hierarchical Clustering) algorithm cluster. This study compares the accuracy values obtained from conventional K- Means with K-Means using AHC. The SSE results obtained on the k-means algorithm using AHC have increased compared to conventional K-Means by 4.8% in 2 clusters and 14.3% in 3 clusters.