1 paper · 2 filters
Thomas Hamm, Ingo Steinwart
We derive improved regression and classification rates for support vector machines using Gaussian kernels under the assumption that the data has some low-dimensional intrinsic stru…