Selecting the Hypothesis Space for Improving the Generalization Ability of Support Vector Machines

被引:0
作者
Anguita, Davide [1 ]
Ghio, Alessandro [1 ]
Oneto, Luca [1 ]
Ridella, Sandro [1 ]
机构
[1] Univ Genoa, Dept Biophys & Elect Engn, I-16145 Genoa, Italy
来源
2011 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN) | 2011年
关键词
CROSS-VALIDATION; CANCER; CLASSIFICATION; DISCOVERY;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The Structural Risk Minimization framework has been recently proposed as a practical method for model selection in Support Vector Machines (SVMs). The main idea is to effectively measure the complexity of the hypothesis space, as defined by the set of possible classifiers, and to use this quantity as a penalty term for guiding the model selection process. Unfortunately, the conventional SVM formulation defines a hypothesis space centered at the origin, which can cause undesired effects on the selection of the optimal classifier. We propose here a more flexible SVM formulation, which addresses this drawback, and describe a practical method for selecting more effective hypothesis spaces, leading to the improvement of the generalization ability of the final classifier.
引用
收藏
页码:1169 / 1176
页数:8
相关论文
共 24 条
  • [1] Alon U., 1999, CELL BIOL
  • [2] On the statistical assessment of classifiers using DNA microarray data
    Ancona, N.
    Maglietta, R.
    Piepoli, A.
    D'Addabbo, A.
    Cotugno, R.
    Savino, M.
    Liuni, S.
    Carella, M.
    Pesole, G.
    Perri, F.
    [J]. BMC BIOINFORMATICS, 2006, 7 (1)
  • [3] Anguita D., 2011, NEUROCOMPUT IN PRESS
  • [4] Anguita D., 2011, INT JOINT C NE UNPUB
  • [5] Anguita D, 2010, IEEE IJCNN
  • [6] A support vector machine with integer parameters
    Angulta, Davide
    Ghio, Alessandro
    Pischiutta, Stefano
    Ridella, Sandro
    [J]. NEUROCOMPUTING, 2008, 72 (1-3) : 480 - 489
  • [7] [Anonymous], 2003, PRACTICAL GUIDE SUPP
  • [8] [Anonymous], 2007, IEEE INT C ICML
  • [9] [Anonymous], 2000, NATURE STAT LEARNING, DOI DOI 10.1007/978-1-4757-3264-1
  • [10] Bartlett PL, 2001, LECT NOTES ARTIF INT, V2111, P224