The problem of multiclass-to-binary reductions in the context of classification with kernel machines continues to attract considerable attention. Indeed, the current understanding of this problem is rather limited. Despite the multitude of proposed solutions no single method is known to be consistently superior to others. We developed a new multi-class classification method that reduces the multi-class problem to a single binary classifier (SBC). Our method constructs the binary problem by embedding smaller binary problems into a single space. We show that the construction of a good embedding, which allows for large margin classification, can be reduced to the task of learning linear combinations of kernels. We observe that a known margin generalization error-bound for standard binary classification applies to our construction. Our empirical examination of the new method indicates that it can outperform one-vs-all, all-pairs and the error-correcting output coding scheme.
By: Ran El-Yaniv; Dmitry Pechyony; Elad Yom-Tov
Published in: H-0243 in 2006
LIMITED DISTRIBUTION NOTICE:
This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.
Questions about this service can be mailed to firstname.lastname@example.org .