PAC-Bayesian Bounds for Randomized Empirical Risk Minimizers
Machine Learning
2009-01-09 v3 Statistics Theory
Statistics Theory
Abstract
The aim of this paper is to generalize the PAC-Bayesian theorems proved by Catoni in the classification setting to more general problems of statistical inference. We show how to control the deviations of the risk of randomized estimators. A particular attention is paid to randomized estimators drawn in a small neighborhood of classical estimators, whose study leads to control the risk of the latter. These results allow to bound the risk of very general estimation procedures, as well as to perform model selection.
Cite
@article{arxiv.0712.1698,
title = {PAC-Bayesian Bounds for Randomized Empirical Risk Minimizers},
author = {Pierre Alquier},
journal= {arXiv preprint arXiv:0712.1698},
year = {2009}
}