A new geometrical hypothesis for clustering and discriminant analysis

  • Jean-Paul Rasson
  • Didier Jacquemin
  • Vincent Bertholet
Conference paper
Part of the Studies in Classification, Data Analysis, and Knowledge Organization book series (STUDIES CLASS)


In this paper, we propose a new clustering method and a new discriminant rule valid on the basic spaced ℜ2. These procedures make use of a new concept in clustering: the concept of closed and connex forms. This hypothesis generalizes the convex hypothesis and is very useful to find non convex natural clusters. Finally, we examine the admissibility of our procedure, in the sense of Fisher and Van Ness [4].


Cluster Analysis Discriminant Analysis Lebesgue Measure Stationary Poisson Point Process Hypervolume Method Replacing the convexity Non Convex case Connex and Closed domains Optimal Closing Admissibility Conditions 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Beaufays, P. and Rasson, J.-P. (1985), A new geometric discriminant rule, in: Computational Statistics Quarterly, 2, 15–30, Elsevier Science B.V.Google Scholar
  2. [2]
    Bock, H. H. (1997), Probability Models for Convex Clusters, in: Classification and Knowledge Organization, Proceedings of the 20th Annual Conference of the Gesellschaft für Klassifikation e.V., University of Freiburg, March 6–8,1996, Klar-Opitz editors, Springer.Google Scholar
  3. [3]
    Cressie, N. A.C. (1991), Statistics for spatial data, John Wiley & Sons, New York.Google Scholar
  4. [4]
    Fisher, L., and Van Ness, J.W. (1971), Admissible clustering procedures, in: Biometrika, 58, 91–104.CrossRefGoogle Scholar
  5. [5]
    Hardy, A., and Rasson, J.-P. (1982), Une nouvelle approche des problèmes de classification automatique, in: Statistiques et Analyse des Données, 7, 41–56.Google Scholar
  6. [6]
    Karr, A.F. (1991), Point processes and their statistical inference, Marcel Dekker, New-York.Google Scholar
  7. [7]
    Krickeberg, K. (1982), Processus ponctuels en statistique, in: Ecole d’été de Probabilités de Saint-FlourX-1980, 205–313, Springer-verlag, Berlin.Google Scholar
  8. [8]
    Matheron, G. (1975), Random Sets and Integral Geometry, John Wiley & sons, New-York.Google Scholar
  9. [9]
    Rasson, J.-P., Hardy, A., Szary, L. and Schindler, P. (1982), An optimal clustering method based on the Rassoris criterion and resulting from a new approach, in: Pattern Recognition Theory and Applications, 63–71, Kitter, Fu and Pau editions.Google Scholar
  10. [10]
    Schmitt, M. and Mattioli, J. (1993), Morphologie mathématique, Logiques Mathématiques Informatiques, Masson, Paris.Google Scholar
  11. [11]
    Serra, J.P. (1982), Image Analysis and Mathematical Morphology, Academic Press, London.Google Scholar
  12. [12]
    Spath, H. (1980), Cluster analysis algorithms for data reduction and classification of objects, Ellis Horwood Limited, Chichester.Google Scholar
  13. [13]
    Stoyan, D., and Stoyan, H. (1994), Fractals, Random Shapes and Point Fields, Methods of Geometrical Statistics, John Wiley & Sons, Chichester.Google Scholar
  14. [14]
    Van Ness, J.W. (1973), Admissible clustering procedures, in: Biometrika, 60, 422–424.Google Scholar

Copyright information

© Springer-Verlag Berlin · Heidelberg 1998

Authors and Affiliations

  • Jean-Paul Rasson
    • 1
  • Didier Jacquemin
    • 1
  • Vincent Bertholet
    • 1
  1. 1.Department of MathematicsFacultés Universitaires Notre-Dame de la PaixNamurBelgium

Personalised recommendations