Title: Performance analysis for clustering algorithms

Authors: Yu Xue; Binping Zhao; Tinghuai Ma

Addresses: School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing, 210044, China; Jiangsu Engineering Center of Network Monitoring, Nanjing University of Information Science and Technology, Nanjing, 210044, China; Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology, Nanjing, 210044, China ' School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing, 210044, China ' School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing, 210044, China

Abstract: There are lots of algorithms for optimal clustering. The main part of clustering algorithms includes the K-means, fuzzy c-means (FCM) and evolution algorithm. The main purpose of this paper was to research the performance and characteristics of these three types of algorithms. One criteria (clustering validity index), namely TRW, was used in the optimisation of classification and eight real-world datasets (glass, wine, ionosphere, biodegradation, connectionist bench, hill-valley, musk, madelon datasets), whose dimension became higher, were applied. We made a performance analysis and concluded that it was easy of the K-means and FCM to fall into a local minimum, and the hybrid algorithm was found more reliable and more efficient, especially on difficult tasks with high dimension.

Keywords: optimal clustering; K-means clustering; fuzzy C-means; FCM clustering; hybrid differential evolution; performance analysis; high dimension; performance evaluation; clustering algorithms.

DOI: 10.1504/IJCSM.2016.080089

International Journal of Computing Science and Mathematics, 2016 Vol.7 No.5, pp.485 - 493

Received: 26 May 2016
Accepted: 08 Aug 2016

Published online: 01 Nov 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article