NeuroCOLT

Neural Networks and Computational Learning Theory

 

About NeuroCOLT

Papers Archive

1994 1995
1996 1997
1998 1999
2000 2001

Books

info@neurocolt.org

NeuroCOLT Technical Report NC-TR-01-097


2001-097
Marginal Boosting

Gunnar Ratsch
Manfred K. Warmuth

ABSTRACT
AdaBoost produces a linear combination of weak hypotheses. It has been observed in practice that the generalization error of the algorithm continues to improve even after all examples are
classified correctly by the current linear combination, i.e.\ by a \emph{hyperplane} in feature space where each weak hypothesis is a dimension. The improvement is attributed to the experimental observation that the distances (margins) of the examples to the separating hyperplane are increasing even when the training error is already zero, i.e.\ all examples are on the correct side of the hyperplane. We give an iterative version of AdaBoost that explicitly maximizes the minimum margin of the examples. We bound the number of iterations and the number of hypotheses used in the final linear combination which approximates the maximum margin hyperplane with a certain precision. This result is shown to be independent from the
size of the hypothesis class -- even infinite hypothesis classes are allowed.

Download PDF