Unified Binary and Multiclass Margin-Based Classification

被引:0
|
作者
Wang, Yutong [1 ,2 ]
Scott, Clayton [1 ,3 ]
机构
[1] Univ Michigan, Dept Elect Engn & Comp Sci, Ann Arbor, MI 48109 USA
[2] Univ Michigan, Michigan Inst Data Sci, Ann Arbor, MI 48109 USA
[3] Univ Michigan, Dept Stat, Ann Arbor, MI 48109 USA
基金
美国国家科学基金会;
关键词
Classification; loss functions; consistency; margins; label encodings; ALGORITHMS;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The notion of margin loss has been central to the development and analysis of algorithms for binary classification. To date, however, there remains no consensus as to the analogue of the margin loss for multiclass classification. In this work, we show that a broad range of multiclass loss functions, including many popular ones, can be expressed in the relative margin form, a generalization of the margin form of binary losses. The relative margin form is broadly useful for understanding and analyzing multiclass losses as shown by our prior work (Wang and Scott, 2020, 2021). To further demonstrate the utility of this way of expressing multiclass losses, we use it to extend the seminal result of Bartlett et al. (2006) on classification-calibration of binary margin losses to multiclass. We then analyze the class of Fenchel-Young losses, and expand the set of these losses that are known to be classification-calibrated.
引用
收藏
页码:1 / 51
页数:51
相关论文
共 50 条
  • [1] Margin-Based Active Learning of Multiclass Classifiers
    Bressan, Marco
    Cesa-Bianchi, Nicolo
    Lattanzi, Silvio
    Paudice, Andrea
    [J]. JOURNAL OF MACHINE LEARNING RESEARCH, 2024, 25
  • [2] A margin-based multiclass generalization bound via geometric complexity
    Munn, Michael
    Dherin, Benoit
    Gonzalvo, Javier
    [J]. TOPOLOGICAL, ALGEBRAIC AND GEOMETRIC LEARNING WORKSHOPS 2023, VOL 221, 2023, 221
  • [3] A note on margin-based loss functions in classification
    Lin, Y
    [J]. STATISTICS & PROBABILITY LETTERS, 2004, 68 (01) : 73 - 82
  • [4] ASYMPTOTIC BEHAVIOR OF MARGIN-BASED CLASSIFICATION METHODS
    Huang, Hanwen
    [J]. 2018 IEEE STATISTICAL SIGNAL PROCESSING WORKSHOP (SSP), 2018, : 463 - 467
  • [5] Efficient margin-based query learning on action classification
    Shimosaka, Masamichi
    Mori, Taketoshi
    Sato, Tomomasa
    [J]. 2006 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS, VOLS 1-12, 2006, : 2778 - +
  • [6] Analysis of The Margin Setting Algorithm as A Margin-based Spherical Classification Method
    Wang, Yi
    Pan, W. David
    Fu, Jian
    [J]. IEEE SOUTHEASTCON 2015, 2015,
  • [7] Improving Classification with Pairwise Constraints: A Margin-Based Approach
    Nguyen, Nam
    Caruana, Rich
    [J]. MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, PART II, PROCEEDINGS, 2008, 5212 : 113 - 124
  • [8] Convergence rates of generalization errors for margin-based classification
    Park, Changyi
    [J]. JOURNAL OF STATISTICAL PLANNING AND INFERENCE, 2009, 139 (08) : 2543 - 2551
  • [9] FAST MARGIN-BASED COST-SENSITIVE CLASSIFICATION
    Nan, Feng
    Wang, Joseph
    Trapeznikov, Kirill
    Saligrama, Venkatesh
    [J]. 2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2014,
  • [10] Data-dependent margin-based generalization bounds for classification
    Kégl, B
    Linder, T
    Lugosi, G
    [J]. COMPUTATIONAL LEARNING THEORY, PROCEEDINGS, 2001, 2111 : 368 - 384