Advertisement

Evolving Principal Clusters: Theory and Application to Management Monitoring

  • Y. Schektman
  • A. Ibrahim
Conference paper

Summary

Principal Clusters (PC) and Evolving Principal Clusters (EPC) constitute applications of inner products (distances) with relationship effects and the general theory of symmetrical (CARS) and dissymmetrical (CARDS) relationship association indices defined in Schektman (1978, 1987).

PC are obtained by maximizing CARS or CARDS, i.e., inertia of well design configurations of points in a well chosen euclidean space. By proving intermediate PC, EPC simulates evolution between two PC obtained on the same population. So EPC provides a “measure” of mobility for individuals which change classes.

Applications of EPC to manage a population of 41 restaurants is presented; restaurants to be monitored are pointed out with the help of a friendly program running on IBM PC.

Keywords

Orthogonal Projection Canonical Correlation Relationship Effect Projection Matrix Association Index 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Croquette A (1980) Quelques Resultats Synthetiques en Analyse de Donnees Multidimensionnelles: Optimalite et Metriques a Effets Relationnels. These de 3eme cycle, Universite Paul SABATIER (U.P.S.), ToulouseGoogle Scholar
  2. Fabre C (1986) Contribution a la Protection des Methodes Relationnelles. These de 3eme cycle, U.P.S.Google Scholar
  3. Goodman LA, Kruskal WH (1954) Measures of Association for Cross Classifications. J.A.S.A. 49: 732–764Google Scholar
  4. Grau D (1983) Mesures des Effets Relationnels, Applications. These de 3eme cycle, U.P.S.Google Scholar
  5. Ibrahim A (1982) Quelques Applications des Analyses en Composantes Principales sous Contraintes et des Metriques a Effets Relationnels en Analyse des Donnees. These de 3eme cycle, U.P.S.Google Scholar
  6. Ibrahim A, Schektman Y (1986) Principal Cluster Analyses, in: Gaul W, Schader M (eds.) Classification as a Tool of Research, Elsever Science Publisher B.V., North Holland, pp 217–223Google Scholar
  7. Schektman Y (1978) Contribution a la Mesure en Facteurs dans les Sciences Experimentales et a la Mise en Oeuvre des Calculs Statistiques. These de doctorat d’Etat, U.P.S.Google Scholar
  8. Schektman Y, Ibrahim A, Barouin L (1986) Quelques Applications des Analyses en Partitions Principales. Journees Francophones de l’A.S.U. ( Lille ): 4Google Scholar
  9. Schektman Y, Jockin J, Vielle D, Hamad A, Sadeg B (1987) Bibliotheque Mathematique pour l’Analyse de Donnees (B.M.A.D.). Resume des Communications X IX, J.F.A.S.U., LausanneGoogle Scholar
  10. Schektman Y (1983) Quelques Reflexions sur les Metriques: Une Approche Originale pour la Protection des Methodes Relationnelles et l’etude de la Regression Multidimensionnelle. Note Interne, MLAD-UPSGoogle Scholar
  11. Schektman Y (1984) Metriques et Analyse de Donnees. Quelques Contributions. Communication au 2eme colloque Franco-Sovietique “Statistique et Analyse de Donnees”, ParisGoogle Scholar
  12. Schektman Y (1986) Cours de DEA (Universites Toulouse et Paris Dauphine)Google Scholar
  13. Schektman Y (1987) A General Euclidean Approach for Measuring and Describing Associations between Several Sets of Variables, in: Recent Developments in Clustering and Data Analysis. Proceedings of the 1st French-Japanese Scientific Seminar, Tokyo, pp 31–42Google Scholar
  14. Stewart D, Love W (1968) A General Canonical Correlation Index. Psychol. Bull. 70Google Scholar

Copyright information

© Springer-Verlag Berlin · Heidelberg 1988

Authors and Affiliations

  • Y. Schektman
    • 1
  • A. Ibrahim
    • 1
  1. 1.Aegide, Greco-CNRS 59 Universite Toulouse Le MIRAILToulouse CedexFrance

Personalised recommendations