Authors: Nikos Pelekis, Dimitris K. Iakovidis, Evangelos E. Kotsifakos, Ioannis Kopanakis
Addresses: Department of Informatics, University of Piraeus, Piraeus, Greece. ' Department of Methodology, History and Theory of Science, University of Athens, Athens, Greece. ' Department of Informatics, University of Piraeus, Piraeus, Greece. ' Technological Educational Institute of Crete, Heraklion, Crete, Greece
Abstract: Challenged by real-world clustering problems this paper proposes a novel fuzzy clustering scheme of datasets produced in the context of intuitionistic fuzzy set theory. More specifically, we introduce a variant of the Fuzzy C-Means (FCM) clustering algorithm that copes with uncertainty and a similarity measure between intuitionistic fuzzy sets, which is appropriately integrated in the clustering algorithm. We describe an intuitionistic fuzzification of colour digital images upon which we applied the proposed scheme. The experimental evaluation of the proposed scheme shows that it can be more efficient and more effective than the well-established FCM algorithm, opening perspectives for various applications.
Keywords: fuzzy clustering; fuzzy C-means; FCM; intuitionistic fuzzy sets; intuitionistic similarity metrics; uncertainty; similarity measures; colour digital images.
International Journal of Business Intelligence and Data Mining, 2008 Vol.3 No.1, pp.45 - 65
Available online: 25 Apr 2008 *Full-text access for editors Access for subscribers Purchase this article Comment on this article