Reducing Number of Classifiers in DAGSVM Based on Class Similarity

  • Marcin Luckner
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6978)


Support Vector Machines are excellent binary classifiers. In case of multi–class classification problems individual classifiers can be collected into a directed acyclic graph structure DAGSVM. Such structure implements One-Against-One strategy. In this strategy a split is created for each pair of classes, but, because of hierarchical structure, only a part of them is used in the single classification process.

The number of classifiers may be reduced if their classification tasks will be changed from separation of individual classes into separation of groups of classes. The proposed method is based on the similarity of classes. For near classes the structure of DAG stays immutable. For the distant classes more than one is separated with a single classifier. This solution reduces the classification cost. At the same time the recognition accuracy is not reduced in a significant way. Moreover, a number of SV, which influences on the learning time will not grow rapidly.


Classification Directed Acyclic Graph Support Vector Machines One–Against–One 


  1. 1.
    Abe, S.: Support Vector Machines for Pattern Classification (Advances in Pattern Recognition). Springer-Verlag New York, Inc., Secaucus (2005)Google Scholar
  2. 2.
    Kumar, M.A., Gopal, M.: A comparison study on multiple binary-class SVM methods for unilabel text categorization. Pattern Recogn. Lett. 32, 9160:311–9160:323 (2010),
  3. 3.
    Bennett, K.P.: Combining support vector and mathematical programming methods for classification, pp. 307–326. MIT Press, Cambridge (1999)Google Scholar
  4. 4.
    Crammer, K., Singer, Y.: On the learnability and design of output codes for multiclass problems. In: COLT 2000: Proceedings of the Thirteenth Annual Conference on Computational Learning Theory, pp. 35–46. Morgan Kaufmann Publishers Inc., San Francisco (2000)Google Scholar
  5. 5.
    Kijsirikul, B., Ussivakul, N., Meknavin, S.: Adaptive directed acyclic graphs for multiclass classification. In: Ishizuka, M., Sattar, A. (eds.) PRICAI 2002. LNCS (LNAI), vol. 2417, pp. 158–168. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  6. 6.
    Kim, H.-C., Pang, S., Je, H.-M., Kim, D., Bang, S.Y.: Constructing support vector machine ensemble. Pattern Recognition 36(12), 2757–2767 (2003)CrossRefzbMATHGoogle Scholar
  7. 7.
    Kressel, U.H.G.: Pairwise classification and support vector machines, pp. 255–268. MIT Press, Cambridge (1999)Google Scholar
  8. 8.
    Platt, J., Cristianini, N., ShaweTaylor, J.: Large margin dags for multiclass classification. In: Solla, S.A., Leen, T.K., Mueller, K.-R. (eds.) Advances in Neural Information Processing Systems 12, pp. 547–553 (2000)Google Scholar
  9. 9.
    Romero, R., Touretzky, D., Thibadeau, R.: Optical chinese character recognition using probabilistic neural networks. Pattern Recognition 8, 1279–1292 (1997)CrossRefGoogle Scholar
  10. 10.
    Stapor, K.: Geographic map image interpretation - survey and problems. Machine Graphics & Vision 9(1/2), 497–518 (2000)Google Scholar
  11. 11.
    Vapnik, V.: The Nature of Statistical Learning Theory. Springer, Heidelberg (1995)CrossRefzbMATHGoogle Scholar
  12. 12.
    Wilson, R.J.: Introduction to graph theory. John Wiley & Sons, Inc., New York (1986)Google Scholar
  13. 13.
    Ye, W., Shang-Teng, H.: Reducing the number of sub-classifiers for pairwise multi-category support vector machines. Pattern Recogn. Lett. 28, 2088–2093 (2007)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Marcin Luckner
    • 1
  1. 1.Faculty of Mathematics and Information ScienceWarsaw University of TechnologyWarsawPoland

Personalised recommendations