Very Fast Kernel SVM under Budget Constraints

David Picard

In this paper we propose a fast online Kernel SVM algorithm under tight budget constraints. We propose to split the input space using LVQ and train a Kernel SVM in each cluster. To…