Grey wolf optimization based parameter selection for support vector machines
ISSN: 0332-1649
Article publication date: 5 September 2016
Abstract
Purpose
The purpose of this paper is to tune support vector machine (SVM) classifier using grey wolf optimizer (GWO).
Design/methodology/approach
The schema of the work aims at extracting the features from the collected data followed by a SVM classifier and metaheuristic optimization to tune the classifier parameters.
Findings
The optimal tuning of classifier parameters lowers errors due to manual elucidation and decreases the risk in human perceptions and repeated visual dignosis.
Originality/value
A novel, GWO based tuning algorithm is used for SVM classifier, which is implemented in classifying the complex and nonlinear biomedical signals like intracranial electroencephalogram.
Keywords
Acknowledgements
The authors are thankful to R.G. Andrzejak, Universitat Pompeu Fabra, Barcelona for providing his research findings along with iEEG signals online (http://ntsa.upf.edu/downloads) with clear descriptions about hardware specifications of data acquisition system and indications on signal for focal and non-focal areas, which helped the authors in successful implementation of the algorithms.
Citation
Eswaramoorthy, S., Sivakumaran, N. and Sekaran, S. (2016), "Grey wolf optimization based parameter selection for support vector machines", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 35 No. 5, pp. 1513-1523. https://doi.org/10.1108/COMPEL-09-2015-0337
Publisher
:Emerald Group Publishing Limited
Copyright © 2016, Emerald Group Publishing Limited