Suboptimal behaviour of Bayes and MDL in classification under misspecification
Statistics Theory
2007-07-16 v1 Information Theory
Machine Learning
math.IT
Statistics Theory
Abstract
We show that forms of Bayesian and MDL inference that are often applied to classification problems can be *inconsistent*. This means there exists a learning problem such that for all amounts of data the generalization errors of the MDL classifier and the Bayes classifier relative to the Bayesian posterior both remain bounded away from the smallest achievable generalization error.
Cite
@article{arxiv.math/0406221,
title = {Suboptimal behaviour of Bayes and MDL in classification under misspecification},
author = {Peter Grunwald and John Langford},
journal= {arXiv preprint arXiv:math/0406221},
year = {2007}
}
Comments
This is a slightly longer version of our paper at the COLT (Computational Learning Theory) 2004 Conference, containing two extra pages of discussion of the main results