Risk upper bounds for general ensemble methods with an application to multiclass classification

Abstract

This paper generalizes a pivotal result from the PAC-Bayesian literature — the C-bound— primarily designed for binary classification to the general case of ensemble methods of voters with arbitrary outputs. We provide a generic version of the C-bound, an upper bound over the risk of models expressed as a weighted majority vote that is based on the first and… (More)
DOI: 10.1016/j.neucom.2016.09.016

Cite this paper

@article{Laviolette2017RiskUB, title={Risk upper bounds for general ensemble methods with an application to multiclass classification}, author={François Laviolette and Emilie Morvant and Liva Ralaivola and Jean-Francis Roy}, journal={Neurocomputing}, year={2017}, volume={219}, pages={15-25} }