P. M. Long.
Minimum majority classification and boosting.
Proceedings of the The
Eighteenth National Conference on Artificial Intelligence,
2002. (Paper available in Postscript and
PDF formats. Software also
available.)
Abstract
Motivated by a theoretical analysis of the generalization of
boosting, we examine learning algorithms that work by trying
to fit data using a simple majority vote over a small number
of a collection of hypotheses. We provide experimental
evidence that an algorithm based on this principle outputs hypotheses
that often generalize nearly as well as those output
by boosting, and sometimes better. We also provide experimental
evidence for an additional reason that boosting algorithms
generalize well, that they take advantage of cases in
which there are many simple hypotheses with independent
errors.