Simple view
Full metadata view
Authors
Statistics
Combining one-versus-one and one-versus-all strategies to improve multiclass SVM classifier
Support Vector Machine (SVM) is a binary classifier, but most of the problems we find in the real-life applications are multiclass. There are many methods of decomposition such a task into the set of smaller classification problems involving two classes only. Two of the widely known are one-versus-one and one-versus-rest strategies. There are several papers dealing with these methods, improving and comparing them. In this paper, we try to combine theses strategies to exploit their strong aspects to achieve better performance. As the performance we understand both recognition ratio and the speed of the proposed algorithm. We used SVM classifier on several different databases to test our solution. The results show that we obtain better recognition ratio on all tested databases. Moreover, the proposed method turns out to be much more efficient than the original one-versus-one strategy.
dc.abstract.en | Support Vector Machine (SVM) is a binary classifier, but most of the problems we find in the real-life applications are multiclass. There are many methods of decomposition such a task into the set of smaller classification problems involving two classes only. Two of the widely known are one-versus-one and one-versus-rest strategies. There are several papers dealing with these methods, improving and comparing them. In this paper, we try to combine theses strategies to exploit their strong aspects to achieve better performance. As the performance we understand both recognition ratio and the speed of the proposed algorithm. We used SVM classifier on several different databases to test our solution. The results show that we obtain better recognition ratio on all tested databases. Moreover, the proposed method turns out to be much more efficient than the original one-versus-one strategy. | pl |
dc.affiliation | Wydział Fizyki, Astronomii i Informatyki Stosowanej : Zakład Technologii Gier | pl |
dc.conference | 9th International Conference on Computer Recognition Systems CORES 2015 | |
dc.conference | 9th International Conference on Computer Recognition Systems CORES 2015 | pl |
dc.conference.city | Wrocław | |
dc.conference.country | Polska | |
dc.conference.datefinish | 2015-05-27 | |
dc.conference.datestart | 2015-05-25 | |
dc.conference.indexscopus | true | |
dc.contributor.author | Chmielnicki, Wiesław - 160876 | pl |
dc.contributor.author | Stąpor, Katarzyna | pl |
dc.contributor.editor | Burduk, Robert | pl |
dc.contributor.editor | Jackowski, Konrad | pl |
dc.contributor.editor | Kurzyński, Marek | pl |
dc.contributor.editor | Woźniak, Michał | pl |
dc.contributor.editor | Żołnierek, Andrzej | pl |
dc.date.accessioned | 2016-06-30T13:24:39Z | |
dc.date.available | 2016-06-30T13:24:39Z | |
dc.date.issued | 2016 | pl |
dc.description.conftype | international | pl |
dc.description.physical | 37-45 | pl |
dc.description.publication | 0,5 | pl |
dc.description.series | Advances in Intelligent Systems and Computing | |
dc.description.seriesnumber | 403 | |
dc.identifier.doi | 10.1007/978-3-319-26227-7_4 | pl |
dc.identifier.eisbn | 978-3-319-26227-7 | pl |
dc.identifier.isbn | 978-3-319-26225-3 | pl |
dc.identifier.serieseissn | 2194-5365 | |
dc.identifier.seriesissn | 2194-5357 | |
dc.identifier.uri | http://ruj.uj.edu.pl/xmlui/handle/item/28563 | |
dc.language | eng | pl |
dc.language.container | eng | pl |
dc.pubinfo | Cham : Springer International Publishing | pl |
dc.rights | Dodaję tylko opis bibliograficzny | * |
dc.rights.licence | bez licencji | |
dc.rights.uri | * | |
dc.subtype | ConferenceProceedings | pl |
dc.title | Combining one-versus-one and one-versus-all strategies to improve multiclass SVM classifier | pl |
dc.title.container | Proceedings of the 9th International Conference on Computer Recognition Systems CORES 2015 | pl |
dc.type | BookSection | pl |
dspace.entity.type | Publication |