Lepskii Principle in Supervised Learning
Statistics Theory
2019-05-28 v1 Statistics Theory
Abstract
In the setting of supervised learning using reproducing kernel methods, we propose a data-dependent regularization parameter selection rule that is adaptive to the unknown regularity of the target function and is optimal both for the least-square (prediction) error and for the reproducing kernel Hilbert space (reconstruction) norm error. It is based on a modified Lepskii balancing principle using a varying family of norms.
Cite
@article{arxiv.1905.10764,
title = {Lepskii Principle in Supervised Learning},
author = {Gilles Blanchard and Peter Mathé and Nicole Mücke},
journal= {arXiv preprint arXiv:1905.10764},
year = {2019}
}