Robust support vector regression networks for function approximation with outliers
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
In order to better restrain the end effects in Hilbert-Huang Transform, support vector regression machines (SVRM), which have the superiority in the time series prediction, are adopted to extend the data at the both ends. In the application of SVRM, the parameters have a great influence on the performance of generalization. In this paper the influence of parameters is discussed, and then an adaptive support vector regression machine is proposed based on the particle swarm optimization (PSO) algorithm. With the parameters optimized by PSO, SVRM can be characterized as self-adaptive and high generalization performance in applications. Experiments show that this method can solve the problem of selecting parameters properly. Contrast to the neural networks methods and HHTDPS designed by Huang et al., end effects can be restrained better and the Intrinsic Mode Functions have less distortion.