On Using Asymmetry Information for Classification in Extended Dissimilarity Spaces

  • Yenisel Plasencia-Calaña
  • Edel B. García-Reyes
  • Robert P. W. Duin
  • Mauricio Orozco-Alzate
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7441)

Abstract

When asymmetric dissimilarity measures arise, asymmetry correction methods such as averaging are used in order to make the matrix symmetric. This is usually needed for the application of pattern recognition procedures, but in this way the asymmetry information is lost. In this paper we present a new approach to make use of the asymmetry information in dissimilarity spaces. We show that taking into account the asymmetry information improves classification accuracy when a small number of prototypes is used to create an extended asymmetric dissimilarity space. If the degree of asymmetry is higher, improvements in classification accuracy are also higher. The symmetrization by averaging also works well in general, but decreases performance for highly asymmetric data.

Keywords

asymmetric dissimilarity dissimilarity spaces prototype selection extended dissimilarity spaces 

References

  1. 1.
    Pekalska, E., Duin, R.P.W.: The Dissimilarity Representation for Pattern Recognition: Foundations and Applications (Machine Perception and Artificial Intelligence). World Scientific Publishing Co., Inc., River Edge (2005)MATHGoogle Scholar
  2. 2.
    Pekalska, E., Duin, R.P.W., Paclík, P.: Prototype selection for dissimilarity-based classifiers. Pattern Recogn. 39(2), 189–208 (2006)MATHCrossRefGoogle Scholar
  3. 3.
    Bunke, H., Riesen, K.: Graph Classification Based on Dissimilarity Space Embedding. In: da Vitoria Lobo, N., Kasparis, T., Roli, F., Kwok, J.T., Georgiopoulos, M., Anagnostopoulos, G.C., Loog, M. (eds.) S+SSPR 2008. LNCS, vol. 5342, pp. 996–1007. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  4. 4.
    Orozco-Alzate, M., Duin, R.P.W., Castellanos-Domínguez, G.: A generalization of dissimilarity representations using feature lines and feature planes. Pattern Recogn. Lett. 30(3), 242–254 (2009)CrossRefGoogle Scholar
  5. 5.
    Pekalska, E., Duin, R.P.W.: Beyond traditional kernels: Classification in two dissimilarity-based representation spaces. IEEE Trans. Syst. Man Cybern. C, Appl. Rev. 38(6), 729–744 (2008)CrossRefGoogle Scholar
  6. 6.
    Pekalska, E., Paclik, P., Duin, R.P.W.: A generalized kernel approach to dissimilarity-based classification. J. Mach. Learn. Res. 2, 175–211 (2002)MathSciNetMATHGoogle Scholar
  7. 7.
    Duin, R.P.W., Pękalska, E.: Non-Euclidean Dissimilarities: Causes and Informativeness. In: Hancock, E.R., Wilson, R.C., Windeatt, T., Ulusoy, I., Escolano, F. (eds.) SSPR & SPR 2010. LNCS, vol. 6218, pp. 324–333. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  8. 8.
    Schölkopf, B., Mika, S., Burges, C.J.C., Knirsch, P., Müller, K.R., Rätsch, G., Smola, A.J.: Input space versus feature space in kernel-based methods. IEEE Trans. Neural Netw. 10(5), 1000–1017 (1999)CrossRefGoogle Scholar
  9. 9.
    Okada, A., Imaizumi, T.: Nonmetric multidimensional scaling of asymmetric proximities. Behaviormetrika 14(21), 81–96 (1987)CrossRefGoogle Scholar
  10. 10.
    Martín-Merino, M., Muñoz, A.: Self Organizing Map and Sammon Mapping for Asymmetric Proximities. In: Dorffner, G., Bischof, H., Hornik, K. (eds.) ICANN 2001. LNCS, vol. 2130, pp. 429–435. Springer, Heidelberg (2001)CrossRefGoogle Scholar
  11. 11.
    Muñoz, A., de Diego, I.M., Moguerza, J.M.: Support vector machine classifiers for asymmetric proximities. In: Kaynak, O., Alpaydın, E., Oja, E., Xu, L. (eds.) ICANN 2003 and ICONIP 2003. LNCS, vol. 2714, pp. 217–224. Springer, Heidelberg (2003)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Yenisel Plasencia-Calaña
    • 1
    • 2
  • Edel B. García-Reyes
    • 1
  • Robert P. W. Duin
    • 2
  • Mauricio Orozco-Alzate
    • 3
  1. 1.Advanced Technologies Application CenterHavanaCuba
  2. 2.Faculty of Electrical Engineering, Mathematics and Computer SciencesDelft University of TechnologyThe Netherlands
  3. 3.Departamento de Informática y ComputaciónUniversidad Nacional de Colombia - Sede ManizalesManizalesColombia

Personalised recommendations