Mean Field Method of Least Square Support Vector Machine for Function Approximation
Wenhui Chen, Changchun Zhu, Wanzhao Cui · Proceedings · 2008
This paper presents a simple mean field method of least square support vector machine for function approximation, which corresponds to a mean field method for the regularization network. The proposed method is illustrated for the Gaussian kernel. A number of simulations are presented to demonstrate the performance of the proposed method including modeling nonlinear function, Mackey-Glass time series and stock prices.