n the Complexity of Good Samples for Learning
Proc. Tenth International Computing and Combinatorics Conference(COCOON 2004), Jeju Island, Korea
Volume Vol. 3106 of Lecture Notes in Computer Science, Springer-Verlag,
In machine-learning, maximizing the sample margin can reduce
the learning generalization-error. Thus samples on which
the target function has a large margin ($\gamma$) convey more information
so we expect
fewer such samples.
In this paper, we estimate the complexity of a class
of sets of large-margin samples
for a general learning problem over a finite domain.
We obtain an explicit dependence of this complexity on $\gamma$ and the sample size.