Penggunaan Metode Rank Order Centroid dalam Penentuan Nilai Centroid (Studi Kasus : Dataset Biji Gandum)


  • ricky dwi saputro Saputro Universitas Sahid Surakarta, Indonesia
  • Dwi Retnoningsih Universitas Sahid Surakarta, Indonesia
  • Hardika Khusnuliawati Universitas Sahid Surakarta, Indonesia


Clustering, K-Means, centroid, ROC


Clustering is a data processing in data mining, namely a clustering process by separating a set of data into smaller groups or clusters based on similar characteristics. A clustering method in data mining is K-Means. The KMeans method processes data into the desired number of clusters and the data will be placed into clusters based on the proximity of the centroid or distance to each cluster. The conventional K-Means method has the disadvantage, namely the initial centroid selection is random so it can produce different results. Therefore, initial centroid selection can improve accuracy in the clustering process. Determining the initial centroid can be done in various ways such as Rank Order Centroid (ROC). ROC is a method that gives a score to each criterion according to its ranking appropriate to score based on its priority level. This research compares the evaluation results with the Davies Bouldin Index (DBI) method, which is a cluster validation method for grouping methods. The results of the clustering method using wheat grain data show that a DBI KMeans value using ROC of 0.33350 with 4 iterations in 2 clusters.

Keywords : Clustering, K-Means, centroid, ROC




Citation Check