1 paper · 1 filter
Thomas Hamm, Ingo Steinwart
We prove minimax optimal learning rates for kernel ridge regression, resp.~support vector machines based on a data dependent partition of the input space, where the dependence of t…