Probability estimates for multi-class classification by pairwise coupling

被引：0
|
作者：
Wu, Ting-Fan [1 ]
Lin, Chih-Jen [1 ]
Weng, Ruby C. [2 ]
机构：
[1] Department of Computer Science, National Taiwan University, Taipei,106, Taiwan
[2] Department of Statistics, National Chengchi University, Taipei,116, Taiwan
来源：
Journal of Machine Learning Research | 2004年 / 5卷
关键词：
Classification (of information) - Probability - Learning systems - Classifiers - Support vector machines - Decision trees;
D O I：
暂无
中图分类号：
学科分类号：
摘要：
Pairwise coupling is a popular multi-class classification method that combines all comparisons for each pair of classes. This paper presents two approaches for obtaining class probabilities. Both methods can be reduced to linear systems and are easy to implement. We show conceptually and experimentally that the proposed approaches are more stable than the two existing popular methods: voting and the method by Hastie and Tibshirani (1998). © 2004 Ting-Fan Wu, Chih-Jen Lin, and Ruby C. Weng.
引用
收藏
页码：975 / 1005
相关论文
共 50 条
  • [1] Probability estimates for multi-class classification by pairwise coupling
    Wu, TF
    Lin, CJ
    Weng, RC
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 16, 2004, 16 : 529 - 536
  • [2] Probability estimates for multi-class classification by pairwise coupling
    Wu, TF
    Lin, CJ
    Weng, RC
    JOURNAL OF MACHINE LEARNING RESEARCH, 2004, 5 : 975 - 1005
  • [3] Multi-class SVM classifier based on pairwise coupling
    Li, ZY
    Tang, SW
    Yan, SC
    PATTERN RECOGNITON WITH SUPPORT VECTOR MACHINES, PROCEEDINGS, 2002, 2388 : 321 - 333
  • [4] A transitivity analysis of bipartite rankings in pairwise multi-class classification
    Waegeman, Willem
    De Baets, Bernard
    INFORMATION SCIENCES, 2010, 180 (21) : 4099 - 4117
  • [5] Probability estimation for multi-class classification using AdaBoost
    Nie, Qingfeng
    Jin, Lizuo
    Fei, Shumin
    PATTERN RECOGNITION, 2014, 47 (12) : 3931 - 3940
  • [6] Nested Dichotomies with probability sets for multi-class classification
    Yang Gen
    Destercke, Sebastien
    Masson, Marie-Helene
    21ST EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE (ECAI 2014), 2014, 263 : 363 - +
  • [7] An improved method to the SVM multi-class classifier based on pairwise coupling
    Chen, Y
    Zhang, GJ
    2003 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS, VOLS 1-5, PROCEEDINGS, 2003, : 3212 - 3215
  • [8] Reliable Multi-class Classification based on Pairwise Epistemic and Aleatoric Uncertainty
    Nguyen, Vu-Linh
    Destercke, Sebastien
    Masson, Marie-Helene
    Huellermeier, Eyke
    PROCEEDINGS OF THE TWENTY-SEVENTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, 2018, : 5089 - 5095
  • [9] Generalized Bradley-Terry models and multi-class probability estimates
    Huang, TK
    Weng, RC
    Lin, CJ
    JOURNAL OF MACHINE LEARNING RESEARCH, 2006, 7 : 85 - 115
  • [10] Feature selection for multi-class classification using pairwise class discriminatory measure and covering concept
    Ji, H
    Bang, SY
    ELECTRONICS LETTERS, 2000, 36 (06) : 524 - 525
12345