Mendelson, ShaharNeeman, Joseph2016-02-112016-02-110090-5364http://hdl.handle.net/1885/733712379Under mild assumptions on the kernel, we obtain the best known error rates in a regularized learning scenario taking place in the corresponding reproducing kernel Hilbert space (RKHS). The main novelty in the analysis is a proof that one can use a regularization term that grows significantly slower than the standard quadratic growth in the RKHS norm.Supported in part by Australian Research Council Discovery Grant DP0559465 and by Israel Science Foundation Grant 666/06.© Institute of Mathematical Statistics, 2002. http://www.sherpa.ac.uk/romeo/issn/0090-5364..."author can archive publisher's version/PDF. On author's personal website or open access repository" from SHERPA/RoMEO site (as at 11/02/16)Regressionreproducing kernel Hilbert spaceregulationleast-squaresmodel selection201010.1214/09-AOS7282016-02-24