Growing kernel-based self-organized maps trained with supervised bias

Article type: Research Article

Authors: Papadimitriou, Stergios | Terzidis, Konstantinos

Affiliations: Department of Information Management, Technological Educational Institute of Kavala, 65404 Kavala, Greece. E-mail: [email protected], [email protected]

Abstract: Most unsupervised learning algorithms ignore prior application knowledge. Also, Self Orgnanized Maps (SOM) based approaches usually develop topographic maps with disjoint and uniform activation regions that correspond to a hard clustering of the patterns at their nodes. We present a novel Self-Organizing map that adapts its parameters in kernel space, grows dynamically up to a size defined with statistical criteria and is capable of incorporating a priori information in the form of a supervised bias at the cluster formation.

Keywords: self-organizing map, reclassification, clustering, multilabeling of data, kernel methods, data mining

DOI: 10.3233/IDA-2004-8202

Journal: Intelligent Data Analysis, vol. 8, no. 2, pp. 111-130, 2004