←
^
→
Computational Learning Theory
Sample Complexity for Infinite H
Alexei Chervonenkis
and
Vladimir Vapnik
So far, we have been using $|H|$ in our equation for sample complexity. Only works for finite $H$.
It can also lead to weak (i.e., too loose/high) bounds.
We now show how to handle infinite $H$ by using the Vapnik-Chervonenkis dimension.
José M. Vidal
.
14 of 26