RBF Kernel Based Support Vector Machine with Universal Approximation and Its Application
The SVM has been used to the nonlinear function mapping successfully, but the universal approximation property of the SVM has never been proved in theory. This paper proves the universal approximation of the SVM with RBF kernel to arbitrary functions on a compact set and deduces it to the approximation of discrete function. From simulation we can see that the RBF kernel based LS-SVM is more effective in nonlinear function estimation and can prevent the system from noise pollution, so it has high generalization ability.
Unable to display preview. Download preview PDF.
- 1.Suykens, J.A.K.: Nonlinear Modeling and Support Vector Machines. In: IEEE Conference on Instrumentation and Measurement Technology, Budapest, Hungary, May, pp. 287–294 (2001)Google Scholar
- 2.Suykens, J.A.K.: Sparse Approximation Using Least Squares Support Vector Machines. In: IEEE Int. Symposium on Circuit and Systems, Geneva, Switzerland, May, pp. 757–760 (2000)Google Scholar
- 4.Wang, J.P., Jing, Z.L.: A Stochastic Fuzzy Neural Network with Universal Approximation and Its Application. In: Proc. of Int. Conf. On Fuzzy Information Processing Theories and Applications, pp. 497–502. Tsinghua University Press & Springer, Beijing (2003)Google Scholar