Authors: David Koloseni; Mario Fedrizzi; Pasi Luukka; Mikael Collan; Jouni Lampinen
Addresses: Laboratory of Applied Mathematics, Lappeenranta University of Technology, P.O. Box 20, FI-53851 Lappeenranta, Finland; Department of Mathematics, University of Dar es salaam, P.O. Box 35062 Dar es salaam, Tanzania ' Department of Industrial Engineering, University of Trento, Via Mesiano 77, I-38123 Trento, Italy ' School of Business, Lappeenranta University of Technology, P.O. Box 20, FI-53851 Lappeenranta, Finland ' School of Business, Lappeenranta University of Technology, P.O. Box 20, FI-53851 Lappeenranta, Finland ' Department of Computer Science, University of Vaasa, P.O. Box 700, FI-65101 Vaasa, Finland; Department of Computer Science, VSB-Technical University of Ostrava, 17. listopadu 15, 70833 Ostrava-Poruba, Czech Republic
Abstract: This paper describes a new classification method that uses the differential evolution algorithm to find an optimal parameter selection for classification of objects. Parameters optimised include the distance measure used, distance parameters, class vectors for each class, and OWA-parameters. The distances yielded for each feature by the optimised distance measures are aggregated into an overall distance by using OWA-based multi-distance aggregation. In this paper, OWA-based multi-distances are applied for the first time in a classification problem. The OWA-based multi-distance aggregation method is tested with five possible OWA-based aggregation schemes and for each one of them an optimal value is computed with DE. The method is tested with five datasets and results show improvement in classification accuracy compared to previous literature.
Keywords: differential evolution; object classification; multi-distance aggregation; distance selection optimisation; pool of distances; ordered weighted average; OWA; parameter selection.
International Journal of Mathematics in Operational Research, 2016 Vol.9 No.4, pp.436 - 451
Received: 09 Dec 2014
Accepted: 08 May 2015
Published online: 23 Sep 2016 *