Reduction from Cost-Sensitive Multiclass Classification to One-versus-One Binary Classification

Hsuan-Tien Lin
Proceedings of the Sixth Asian Conference on Machine Learning, PMLR 39:371-386, 2015.

Abstract

Many real-world applications require varying costs for different types of mis-classification errors. Such a cost-sensitive classification setup can be very different from the regular classification one, especially in the multiclass case. Thus, traditional meta-algorithms for regular multiclass classification, such as the popular one-versus-one approach, may not always work well under the cost-sensitive classification setup. In this paper, we extend the one-versus-one approach to the field of cost-sensitive classification. The extension is derived using a rigorous mathematical tool called the cost-transformation technique, and takes the original one-versus-one as a special case. Experimental results demonstrate that the proposed approach can achieve better performance in many cost-sensitive classification scenarios when compared with the original one-versus-one as well as existing cost-sensitive classification algorithms.

Cite this Paper


BibTeX
@InProceedings{pmlr-v39-lin14, title = {Reduction from Cost-Sensitive Multiclass Classification to One-versus-One Binary Classification}, author = {Lin, Hsuan-Tien}, booktitle = {Proceedings of the Sixth Asian Conference on Machine Learning}, pages = {371--386}, year = {2015}, editor = {Phung, Dinh and Li, Hang}, volume = {39}, series = {Proceedings of Machine Learning Research}, address = {Nha Trang City, Vietnam}, month = {26--28 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v39/lin14.pdf}, url = {https://proceedings.mlr.press/v39/lin14.html}, abstract = {Many real-world applications require varying costs for different types of mis-classification errors. Such a cost-sensitive classification setup can be very different from the regular classification one, especially in the multiclass case. Thus, traditional meta-algorithms for regular multiclass classification, such as the popular one-versus-one approach, may not always work well under the cost-sensitive classification setup. In this paper, we extend the one-versus-one approach to the field of cost-sensitive classification. The extension is derived using a rigorous mathematical tool called the cost-transformation technique, and takes the original one-versus-one as a special case. Experimental results demonstrate that the proposed approach can achieve better performance in many cost-sensitive classification scenarios when compared with the original one-versus-one as well as existing cost-sensitive classification algorithms.} }
Endnote
%0 Conference Paper %T Reduction from Cost-Sensitive Multiclass Classification to One-versus-One Binary Classification %A Hsuan-Tien Lin %B Proceedings of the Sixth Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Dinh Phung %E Hang Li %F pmlr-v39-lin14 %I PMLR %P 371--386 %U https://proceedings.mlr.press/v39/lin14.html %V 39 %X Many real-world applications require varying costs for different types of mis-classification errors. Such a cost-sensitive classification setup can be very different from the regular classification one, especially in the multiclass case. Thus, traditional meta-algorithms for regular multiclass classification, such as the popular one-versus-one approach, may not always work well under the cost-sensitive classification setup. In this paper, we extend the one-versus-one approach to the field of cost-sensitive classification. The extension is derived using a rigorous mathematical tool called the cost-transformation technique, and takes the original one-versus-one as a special case. Experimental results demonstrate that the proposed approach can achieve better performance in many cost-sensitive classification scenarios when compared with the original one-versus-one as well as existing cost-sensitive classification algorithms.
RIS
TY - CPAPER TI - Reduction from Cost-Sensitive Multiclass Classification to One-versus-One Binary Classification AU - Hsuan-Tien Lin BT - Proceedings of the Sixth Asian Conference on Machine Learning DA - 2015/02/16 ED - Dinh Phung ED - Hang Li ID - pmlr-v39-lin14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 39 SP - 371 EP - 386 L1 - http://proceedings.mlr.press/v39/lin14.pdf UR - https://proceedings.mlr.press/v39/lin14.html AB - Many real-world applications require varying costs for different types of mis-classification errors. Such a cost-sensitive classification setup can be very different from the regular classification one, especially in the multiclass case. Thus, traditional meta-algorithms for regular multiclass classification, such as the popular one-versus-one approach, may not always work well under the cost-sensitive classification setup. In this paper, we extend the one-versus-one approach to the field of cost-sensitive classification. The extension is derived using a rigorous mathematical tool called the cost-transformation technique, and takes the original one-versus-one as a special case. Experimental results demonstrate that the proposed approach can achieve better performance in many cost-sensitive classification scenarios when compared with the original one-versus-one as well as existing cost-sensitive classification algorithms. ER -
APA
Lin, H.. (2015). Reduction from Cost-Sensitive Multiclass Classification to One-versus-One Binary Classification. Proceedings of the Sixth Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 39:371-386 Available from https://proceedings.mlr.press/v39/lin14.html.

Related Material