Handling concept drifts in incremental learning with support vector machines

Abstract

With the increase in the size of real-world databases, there is an ever-increasing need to scale up inductive learning algorithms.