Let be i.i.d. observations with , , being some compact metric space. The authors consider estimates for the regression function , where , ,
is a continuous bounded function (kernel), is a regularization parameter.
Consistency of is demonstrated under the assumptions that , and the true regression function belongs to the closure of in some suitable reproducing kernel Hilbert space.
To analyze the rates of convergence the authors make assumptions of the form for some , where , . E.g. if then choosing they get .
Results of simulations are presented for and the Gaussian kernel .