Riesz basis-based reproducing kernel and SVM
Deqiang Zhou, Luoqing Li · Journal of Zhejiang University(Sciences Edition) · 2004
The choice of the reproducing kernel is important in SVM technic. A reproducing kernel Hilbert space with a group of Riesz basis in L~2(R) is built, and generalized to widely used orthogonal wavelet. Considering regularized approximation of function in a reproducing kernel Hilbert space based on orthogonal wavelet, some studies on SVM have been done incorporating with wavelet analysis. It is proved that both the discrete approximation and discrete detail of SVM have something to do with support vectors.