The Kohonen algorithm: A powerful tool for analysing and representing multidimensional quantitative and qualitative data

  • Marie Cottrell
  • Patrick Rousset
Methodology for Data Analysis, Task Selection and Nets Design
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1240)


The simultaneous analysis of quantitative and qualitative variables is not an easy task in general. When a linear model is appropriate, the Generalized Linear Models are commonly used with success. But when the intrinsic structure of the data is not at all linear, they give very poor and confusing results. In this paper, we extensively study how to use the (non linear) Kohonen maps to solve some of the interesting problems which are encountered in data analysis: how to realize a rapid and robust classification based on the quantitative variables, how to visualize the classes, their differences and homogeneity, how to cross the classification with the remaining qualitative variables to interpret the classification and put in evidence the most important explanatory variables.


Kohonen maps classification multidimensional data analysis general non linear models neural networks 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    T.W.Anderson, An introduction to multivariate statistical analysis, Wiley, 2nd ed., New York, 1984.Google Scholar
  2. [2]
    F.Blayo P.Demartines, Data analysis: how to compare Kohonen neural networks to other techniques ?, In Proceedings of IWANN'91, Springer, p. 469–476, 1991.Google Scholar
  3. [3]
    M.Cottrell, J.C.Fort, Etude d'un algorithme d'auto-organisation, Annales de l'Institut Poincaré, Vol. 23, 1, 1–20, 1987.Google Scholar
  4. [4]
    M.Cottrell, P.Letremy, E.Roy, E., Analysing a Contingency Table with Kohonen Maps: a Factorial Correspondence Analysis, Proceedings of IWANN'93, Springer Verlag, p. 305–311, 1993.Google Scholar
  5. [5]
    M.Cottrell, J.C.Fort, G.Pagès, Two or three things that we know about the Kohonen algorithm, in Proc of ESANN'94, M. Verleysen Ed., D Facto, Bruxelles, p.235–244, 1995.Google Scholar
  6. [6]
    M.Cottrell, S. Ibbou, Multiple Correspondence Analysis of a crosstabulations matrix using the Kohonen algorithm, in Proc of ESANN'95, M. Verleysen ED, D Facto, Bruxelles, p. 27–32, 1995.Google Scholar
  7. [7]
    M.Cottrell, B.Girard, Y. Girard, C.Muller and P.Rousset, Daily Electrical Power Curves: Classification and Forecasting Using a Kohonen Map, From Natural to Artificial Neural Computation, Proc. IWANN'95, Springer, p. 1107–1113,1995.Google Scholar
  8. [8]
    M.Cottrell, B.Girard, Y.Girard, M.Mangeas, Neural Modeling for Time Series: A Statistical Stepwise Method for Weight Elimination, IEEE Tr. on Neural Networks, Nov. 1995, Vol. 6, No 6, p. 1355–1364, 1995.Google Scholar
  9. [9]
    M.Cottrell, Bodt, A Kohonen Map Representations to Avoid Misleading Interpretations, in Proc of ESANN'96, M. Verleysen Ed., D Facto, Bruxelles, p. 103–110, 1996.Google Scholar
  10. [10]
    A.Dempster, N.Laird, D.Rubin, Maximum likelihood from incomplete data via the EM algorithm, J. Roy Stat. Soc., B39, p. 1–38, 1977.Google Scholar
  11. [11]
    A.Eydoux, D.Wuhl, Traitement statistique et économétrique des données ANPE sur le chômage et les activités réduites, Technical Report, Université Paris 1, 1996.Google Scholar
  12. [12]
    F.Gardes, P.Gaubert, P.Rousset, Cellulage de données d'enquêtes de consommation par une méthode neuronale, Preprint SAMOS #69, 1997.Google Scholar
  13. [13]
    P.Gaubert, S.Ibbou, C.Tutin, Housing market segmentation and price mechanisms in the Parisian metropolis, International Journal of Urban and Regional Research, to appear, 1995.Google Scholar
  14. [14]
    T.Kohonen,. Self-organization and Associative Memory, 3∘ed., Springer, 1993.Google Scholar
  15. [15]
    T.Kohonen, Self Organizing Maps, Springer Series in Information Sciences Vol 30, Springer, 1995.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Marie Cottrell
    • 1
  • Patrick Rousset
    • 1
  1. 1.SAMOS, Université Paris 1Paris Cedex 13France

Personalised recommendations