Show simple item record

dc.contributor.authorErtekin, Seyda
dc.contributor.authorRudin, Cynthia
dc.date.accessioned2012-12-13T20:37:17Z
dc.date.available2012-12-13T20:37:17Z
dc.date.issued2011-10
dc.date.submitted2011-08
dc.identifier.issn1532-4435
dc.identifier.issn1533-7928
dc.identifier.urihttp://hdl.handle.net/1721.1/75726
dc.description.abstractWe demonstrate that there are machine learning algorithms that can achieve success for two separate tasks simultaneously, namely the tasks of classification and bipartite ranking. This means that advantages gained from solving one task can be carried over to the other task, such as the ability to obtain conditional density estimates, and an order-of-magnitude reduction in computational time for training the algorithm. It also means that some algorithms are robust to the choice of evaluation metric used; they can theoretically perform well when performance is measured either by a misclassification error or by a statistic of the ROC curve (such as the area under the curve). Specifically, we provide such an equivalence relationship between a generalization of Freund et al.’s RankBoost algorithm, called the “P-Norm Push,” and a particular cost-sensitive classification algorithm that generalizes AdaBoost, which we call “P-Classification.”We discuss and validate the potential benefits of this equivalence relationship, and perform controlled experiments to understand P-Classification’s empirical performance. There is no established equivalence relationship for logistic regression and its ranking counterpart, so we introduce a logistic-regression-style algorithm that aims in between classification and ranking, and has promising experimental performance with respect to both tasks.en_US
dc.description.sponsorshipNational Science Foundation (U.S.) (Grant no. IIS-1053407)en_US
dc.description.sponsorshipMassachusetts Institute of Technology. Energy Initiativeen_US
dc.language.isoen_US
dc.publisherAssociation for Computing Machineryen_US
dc.relation.isversionofhttp://jmlr.csail.mit.edu/papers/volume12/ertekin11a/ertekin11a.pdfen_US
dc.rightsArticle is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.en_US
dc.sourceMIT Pressen_US
dc.titleOn equivalence relationships between classification and ranking algorithmsen_US
dc.typeArticleen_US
dc.identifier.citationErtekin, Seyda and Cynthia Rudin. "On Equivalence Relationships Between Classification and Ranking Algorithms." Journal of Machine Learning Research 12 (2011) 2905-2929.en_US
dc.contributor.departmentSloan School of Managementen_US
dc.contributor.mitauthorErtekin, Seyda
dc.contributor.mitauthorRudin, Cynthia
dc.relation.journalJournal of Machine Learning Researchen_US
dc.eprint.versionFinal published versionen_US
dc.type.urihttp://purl.org/eprint/type/JournalArticleen_US
eprint.statushttp://purl.org/eprint/status/PeerRevieweden_US
dspace.orderedauthorsErtekin, Seyda; Rudin, Cynthiaen_US
dc.identifier.orcidhttps://orcid.org/0000-0001-6541-1650
mit.licensePUBLISHER_POLICYen_US
mit.metadata.statusComplete


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record