%0 Conference Paper %1 shashua2003ranking %A Shashua, Amnon %A Levin, Anat %B In Proceedings of Advances in Neural Information Processing Systems %D 2003 %E NIPS, %K learning oc ranking sum %T Ranking with Large Margin Principle: Two Approaches %U http://books.nips.cc/papers/files/nips15/AA58.pdf %X We discuss the problem of ranking instances with the use of a “large margin” principle. We introduce two main approaches: the first is the “fixed margin” policy in which the margin of the closest neighboring classes is being maximized — which turns out to be a direct generalization of SVM to ranking learning. The second approach allows for different margins where the sum of margins is maximized. This approach is shown to reduce to \nu-SVM when the number of classes. Both approaches are optimal in size of n where n is the total number of training examples. Experiments performed on visual classification and “collaborative filtering” show that both approaches outperform existing ordinal regression algorithms applied for ranking and multiclass SVM applied to general multiclass classification.

@inproceedings{shashua2003ranking, abstract = {We discuss the problem of ranking instances with the use of a “large margin” principle. We introduce two main approaches: the first is the “fixed margin” policy in which the margin of the closest neighboring classes is being maximized — which turns out to be a direct generalization of SVM to ranking learning. The second approach allows for different margins where the sum of margins is maximized. This approach is shown to reduce to \nu-SVM when the number of classes. Both approaches are optimal in size of n where n is the total number of training examples. Experiments performed on visual classification and “collaborative filtering” show that both approaches outperform existing ordinal regression algorithms applied for ranking and multiclass SVM applied to general multiclass classification. }, added-at = {2012-12-25T14:52:11.000+0100}, author = {Shashua, Amnon and Levin, Anat}, biburl = {https://www.bibsonomy.org/bibtex/284662e39711650b021786f8908bef03c/nosebrain}, booktitle = {In Proceedings of Advances in Neural Information Processing Systems}, description = {Two svm formulations of the ranking or rather ordinal regression are introduced: one which maximizes the individual margins separating the different classes and one which maximizes the sum of the margins. There might be a way to introduce this types of losses in the mmmf formulation.}, editor = {NIPS}, interhash = {20c44188b32fe08a6ea3c5f63731099b}, intrahash = {84662e39711650b021786f8908bef03c}, keywords = {learning oc ranking sum}, timestamp = {2013-03-16T16:01:16.000+0100}, title = {Ranking with Large Margin Principle: Two Approaches}, url = {http://books.nips.cc/papers/files/nips15/AA58.pdf}, year = 2003 }