Abstract

This paper proposes an improved version of particle swarm optimization (PSO) algorithm for the training of a neural network (NN). An architecture for the NN trained by PSO (standard PSO, improved PSO) is also introduced. This architecture has a data preprocessing mechanism which consists of a normalization module and a data-shuffling module. Experimental results showed that the NN trained by improved PSO (IPSO) achieved better performance than both the NN trained by standard PSO and the NN trained by back-propagation (BP) algorithm. The effectiveness concerning the recognition rate and the minimum learning error of the data preprocessing modules (normalization module, data-shuffling module) was also demonstrated through the experiments.

References

• S. Haykin, Neural networks and learning machines, 3rd edn, Prentice Hall, 2008
• R. H. Nielsen, Theory of the backpropagation neural network, In processing of the international conference on neural networks, pp. 693-605, 1989 R. Rojas, Neural networks - a systematic introduction, Springer-Verlag, 1996
• J. R. Zhang, J. Zhang, T. M. Lok. M. R. Lyu, A hybrid particle swarm optimizationback-propagation algorithm for feedforward neural network training, Applied mathematics and computation, vol. 185, pp. 10261037, 2007
• Z.A. Bashir, M.E. El-Hawary, Applying Wavelets to Short- Term Load Forecasting Using PSO-Based Neural Networks, IEEE transactions on power systems, vol. 46, pp. 268-275, 2016
• A. Suresh, K. V. Harish, N. Radhika, Particle Swarm Optimization over Back Propagation Neural Network for Length of Stay Prediction, In processing of the international conference on information and communication technologies, vol. 24, no.1, pp. 20-27, 2009
• V. G. Gudise, G. K. Venayagamoorthy, Comparison of particle swarm optimization and backpropagation as training algorithms for neural networks, In processing of 2003 IEEE swarm intelligence symposium, pp. 110-117, 2003
• M. T. Das, L. C. Dulger, Signature verification (SV) toolbox: Application of PSO-NN, Engineering applications of artificial intelligence, vol. 22, issue 4-5, pp. 688-694, 2009
• R. Mendes, et al., Particle swarms for feedforward neural network training, In processing of the IEEE international joint conference on neural networks, vol.2, pp.1895-1899, 2002
• K. W. Chau, Application of a PSO-based neural network in analysis of outcomes of construction claims. Automation in construction, vol. 16, no. 5, 642-646, 2007
• G.Montavon, G. B. Orr, K. R. Muller Neural networks: tricks of the trade, 2nd edn, Springer, 2012
• J. Kennedy, R. Eberhart, Particle swarm optimization, In processing of the IEEE international conference on neural networks, vol. 4, pp.1942-1948, 1995
• R. Eberhart, Y. Shi, Particle swarm optimization: developments, applications and resources, In processing of the 2001 IEEE international conference on congress on evolutionary computation, vol. 1, pp. 81-86, 2001
• Y. Shi and R. Eberhart, Empirical study of particle swarm optimization, In processing of international conference on evolutionary computation, pp. 1945-1950, 1999
• J. Han, M. Kamber, J. Pei, Data mining: concepts and techniques, 3rd edn, Morgan Kaufmann, 2011
• M. Lichman, UCI Machine Learning Repository, http://archive.ics.uci.edu/ml, accessed Sep. 03 2016
• E. Orhan, A. C. Tanrikulu, A. Abakay, F. Temurtasa, An approach based on probabilistic neural network for diagnosis of Mesotheliomas disease, Comput electr eng, vol. 38, issue 1, pp. 75-81, 2012