| |
Abstract:
We consider online learning in a Reproducing Kernel Hilbert
Space. Our method is computationally efficient and leads to
simple algorithms. In particular we derive update equations for
classification, regression, and novelty detection. The inclusion
of the ν-trick allows us to give a robust parameterization.
Moreover, unlike in batch learning where the ν-trick only
applies to ε-insensitive loss function we are able to
derive general trimmed-mean types of estimators such as for
Huber's robust loss.
|
