Paper 10: Support Vector Machine Series 11: Some theoretical additions to the Kernel ii--nuclear method, about reproducing Kernel Hilbert Space and Representer theorem.

Before we introduced how to use the Kernel method to generalize linear SVM so that it can deal with non-linear situations, the method used here is through a nonlinear mapping .  () The original data is mapped so that the original nonlinear problem becomes a linear problem in the space after mapping. Then we use the kernel function to simplify the calculation, so that the method in practice becomes feasible. However, from the linear to the non-linear generalization we did not deduce the SVM's formula from the beginning, but simply the resulting classification function

Paper 10: Support Vector Machine Series 11: Some theoretical additions to the Kernel ii--nuclear method, about reproducing Kernel Hilbert Space and Representer theorem.