Under 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.
Publié le : 2010-02-15
Classification:
Regression,
reproducing kernel Hilbert space,
regulation,
least-squares,
model selection,
68Q32,
60G99
@article{1262271623,
author = {Mendelson, Shahar and Neeman, Joseph},
title = {Regularization in kernel learning},
journal = {Ann. Statist.},
volume = {38},
number = {1},
year = {2010},
pages = { 526-565},
language = {en},
url = {http://dml.mathdoc.fr/item/1262271623}
}
Mendelson, Shahar; Neeman, Joseph. Regularization in kernel learning. Ann. Statist., Tome 38 (2010) no. 1, pp. 526-565. http://gdmltest.u-ga.fr/item/1262271623/