FDTD time series extrapolation by the least squares support vector machine method with the particle swarm optimization technique

A new combination of particle swarm optimization (PSO) and least-squares support vector machines (LS-SVM) technique for FDTD time series forecasting is presented. In this paper, the PSO is extended to optimize the hyperparameter used in the LS-SVM algorithm. Numerical simulations demonstrate that the PSO method can efficiently get the optimal value of the hyperparameter used in the LS-SVM algorithm. And the PSO/spl I.bar/LS-SVM method can improve the computational efficiency of the FDTD algorithm when compared with the direct FDTD method.

[1]  James Kennedy,et al.  Particle swarm optimization , 2002, Proceedings of ICNN'95 - International Conference on Neural Networks.

[2]  J. Kennedy,et al.  Matching algorithms to problems: an experimental test of the particle swarm and some genetic algorithms on the multimodal problem generator , 1998, 1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98TH8360).

[3]  Johan A. K. Suykens,et al.  Least squares support vector machine classifiers: a large scale algorithm , 1999 .

[4]  Johan A. K. Suykens,et al.  Financial time series prediction using least squares support vector machines within the evidence framework , 2001, IEEE Trans. Neural Networks.

[5]  XiangMo Zhao,et al.  Traffic flow time series prediction based on statistics learning theory , 2002, Proceedings. The IEEE 5th International Conference on Intelligent Transportation Systems.

[6]  Y. Rahmat-Samii,et al.  Particle swarm optimization in electromagnetics , 2004, IEEE Transactions on Antennas and Propagation.

[7]  Xiaowei Sun,et al.  Low-pass filters of wide stop-band based on photonic band-gap structures , 2004 .

[8]  R. S. Chen,et al.  A combination of FDTD and least-squares support vector machines for analysis of microwave integrated circuits , 2005 .