Abstract—Abstract: In this article, we propose a comparison between several validity indices by combining them with the k-means algorithm in order to determine the optimal number of clusters. The method consists at varying the number of clusters in a predefined range [kmin, kmax] and extracts the best indices values representing the final number of clusters. The compared indices are the Davies-Bouldin index, the Sum of Squares index, the Xie-Benie index, the Artur Starczewski index (SRT), the Pakhira-Bandyopadhyay-Maulik Fuzzy index, the Bayesian information Criterion, the Silhouette index and the Wang & al index (WSJ). The experimentation and comparison of the validity indices were performed on synthetic and satellite datasets. The results confirm the effectiveness of some indices such as the Sum of Squares index among a large set of data.

Index Terms—Key words: Clustering, generation of data, index of validity, K-means.

Earth Observation Department, Centre of Space Techniques, Algeria.

Cite:Hanifi Meroufel, Habib Mahi, Nezha Farhi, "Comparative Study between Validity Indices to Obtain the Optimal Cluster," International Journal of Computer and Electrical Engineering vol. 9, no. 1, pp. 343-350, 2017.