Symbolic Pattern Classifiers Based on the Cartesian System Model

  • Manabu Ichino
  • Hiroyuki Yaguchi
Part of the Studies in Classification, Data Analysis, and Knowledge Organization book series (STUDIES CLASS)


As symbolic pattern classifiers, this paper presents region oriented methods based on the Cartesian system model which is a mathematical model to treat symbolic data. Our region oriented methods are able to use locally effective information to discriminate between pattern classes. This fact may achieve, at least superficially, a perfect discrimination of the pattern classes under a finite design set. Therefore, we have to take a ballance between the separability between classes and the generality of class desciptions. We describe this viewpoint theoretically and experimentally in order to assert the importance of feature selection which is essentially important in any pattern classification problem. We present also an example based on symbolic data in order to illustrate the usefulness of our approach.


Feature Selection Feature Space Pattern Classification Euclidean Plane Symbolic Data 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Bow, S. T. (1992): Pattern Recognition and Image Preprocessing, Mercel Dekker.Google Scholar
  2. Diday, E. (1988): The symbolic approach in clustering. In Classification and Related Methods of Data Analysis„ Bock, H. H. (ed. ), Elsevier.Google Scholar
  3. Stoffel, J. C. (1974): A classifier design technique for discrete pattern recognition problems. IEEE Trans. Compt., C-23, pp. 428–441.Google Scholar
  4. Michalski, R. S. (1980): Pattern recognition as rule-guided inductive inference. IEEE Trans. Pattern Anal. and Mach. Intell. PAMI-2, pp. 549–361.Google Scholar
  5. Quinlan, J.R. (1986): Introduction of Decision Tree, Machine Learning, 1, pp. 81106.Google Scholar
  6. Rumelhart, D.E.R. and McClelland (1986): Parallel Distributed Processing, MIT Press.Google Scholar
  7. Ichino, M. (1979): A nonparametric multiclass pattern classifier. IEEE Trans. Syst. Man, Cybern. 9, pp. 345–352.Google Scholar
  8. Ichino, M. (1981): Nonparametric feature selection method based on local interclass structure, IEEE Trans. on Syst. Man. Cybern. 11. pp. 289–296.Google Scholar
  9. Ichino, M and Sklansky, J. (1985): The relative neighborhood graph for mixed feature variables, Pattern Recognition, 18, 2, pp. 161–167.Google Scholar
  10. Ichino, M. (1986): Pattern classification based on the Cartesian join system: A general tool for feature selection, In Proc. IEEE Int. Conf. on SAM (Atlanta).Google Scholar
  11. Ichino, M. (1988): A general pattern classification method for mixed feature problems. Trans IEICE Japan J-71-D, PP. 92–101 (in Japanese).Google Scholar
  12. Ichino, M. (1993): Feature selection for symbolic data classification. In New Ap-proaches in Classification and Data Analysis, Diday, E. et al. (ed.), Springer-Verlag.Google Scholar
  13. Ichino, M and Yaguchi, H. (1995): Generalized Minkowski metrics for mixed feturetype data analysis, IEEE Trans. Syst. Man. Cybern. 24. 4. pp. 698–708.Google Scholar
  14. Ichino, M., Yaguchi, H. and Diday, E. (1995): A fuzzy symbolic pattern classifier. OSDA ‘85, Paris.Google Scholar
  15. Yaguchi, H., Ichino, M. and Diday, E. (1995): A knowledge acquisition system based on the Cartesian space model. OSDA’95, Paris.Google Scholar

Copyright information

© Springer Japan 1998

Authors and Affiliations

  • Manabu Ichino
    • 1
  • Hiroyuki Yaguchi
    • 1
  1. 1.Tokyo Denki UniversityHatoyama, SaitamaJapan

Personalised recommendations