Support vector classification with Gaussian RBF kernel is sensitive to the kernel width. Small kernel width may cause over-fitting, and large one under-fitting. The so-called optimal kernel width is merely selected based on the tradeoff between under-fitting loss and over-fitting loss. So, there exists urgent need to further reduce the tradeoff loss. To circumvent this, we scale the kernel width in a distribution-dependent way. Experiments validate the feasibiity of this method. Existing problems are also discussed.
展开▼