Colonial Competitive Algorithm Assisted Least Squares Support Vector Machines

Abstract:

The use of least squares support vector machine (LSSVM), a novel machine learning method, for classification and function approximation has increased over the past few years especially due to its high generalization performance. However, LSSVM is plagued by the drawback that the hyper-parameters, which largely determine the quality of LSSVM models, have to be defined by the user, and this increases the difficulty of applying LSSVM and limits its use on academic and industrial platforms. In this paper we present a novel method of automatically tuning the hyper-parameters of LSSVM based on colonial competitive algorithm (CCA), a newly developed evolutionary algorithm inspired by imperialistic competition mechanism. To show the efficacy of the CCA assisted LSSVM methodology, we have tested it on several benchmark examples. The study suggests that proposed paradigm can be a competitive and powerful tool for classification and function approximation.

• Periodical: Advanced Materials Research (Volumes 255-260)
• Pages: 2082-2086
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.255-260.2082
• Online since: May 2011
• Authors: Yuan Ren, Guang Chen Bai
• Keywords: Colonial Competitive Algorithm, Machine Learning (ML), Optimization, Parameter Tuning

References

[1] V. Vapnik. The nature of statistical learning theory. New York: Springer-Verlag, 1995.

[2] J. Suykens and J. Vandewalle. Least squares support vector machine classifiers. Neural Processing Letters, 1999, 9(3), 293–300.

[3] R. Rajabioun, E. A. Gargari and C. Lucas. Colonial competitive algorithm as a tool for Nash equilibrium point achievement. LNCS, 2008, Volume 5073: 680–695.

DOI: 10.1007/978-3-540-69848-7_55

[4] L. Zhang, W. Zhou and L. Jiao. Wavelet support vector machine. IEEE Trans Syst Man Cybern Part B Cybern, 2004, 34(1), 34–39.

DOI: 10.1109/tsmcb.2003.811113

[5] K. Ito, R. Nakano. Optimizing support vector regression hyperparameters based on cross-validation. Proceedings of the International Joint Conference on Neural Networks, 2003, Volume 3: 2077–2082.

DOI: 10.1109/ijcnn.2003.1223728