Skip Navigation

Introduction to Kernel Methods for Machine Learning

seis.bristol.ac.uk /~enicgc/pubs/2000/svmintro.pdf

Kernel methods give a systematic and principled approach to training learning machines and the good generalization performance achieved can be readily justified using statistical learning theory or Bayesian arguments. We describe how to use kernel methods for classification, regression and novelty detection and in each case we find that training can be reduced to optimization of a convex cost function.

0
0 comments