Compstat pp 243-248 | Cite as

Comparing Two Partitions: Some Proposals and Experiments

  • Gilbert Saporta
  • Genane Youness

Abstract

We propose a methodology for finding the empirical distribution of the Rand’s measure of association when the two partitions only differ by chance. For that purpose we simulate data coming from a latent profile model and we partition them according to 2 groups of variables. We also study two other indices: the first is based on an adaptation of Mac Nemar’s test, the second being Jaccard’s index. Surprisingly, the distributions of the 3 indices are bimodal.

Keywords

Latent class K-means Rand index Jaccard index partitions 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Bartholomew, D.J. & Knott, M. (1999). Latent Variable Models and Factor Analysis, London: Arnold.MATHGoogle Scholar
  2. Green, P.& Kreiger, A. (1999). A Generalized Rand-Index Method for Consensus Clustering of Separate Partitions of the Same Data Base, Journal of Classification, 16, 63–89.CrossRefGoogle Scholar
  3. Hubert, L. & Arabie, P.(1985). Comparing partitions, Journal of Classification, 2, 193–198.CrossRefGoogle Scholar
  4. Idrissi, A. (2000). Contribution à l’unification de Critères d ‘Association pour Variables Qualitatives, Ph.D., Paris: Université Pierre et Marie Curie.Google Scholar
  5. Marcotorchino, J.F.& El Ayoubi, N. (1991). Paradigme logique des écritures relationnelles de quelques critéres fondamentaux d’association, Revue de Statistique Appliquée, 39, 2, 25–46.MATHGoogle Scholar
  6. Saporta, G. (1997). Problèmes posés par la comparaison de classifications dans des enquêtes différentes, in: Proceedings of the 53rd Session of the International Statistical Institute. Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2002

Authors and Affiliations

  • Gilbert Saporta
    • 1
  • Genane Youness
    • 2
  1. 1.Chaire de Statistique Appliquée-CEDRICCNAMParisFrance
  2. 2.CNAM-ISAEBeirutLebanon

Personalised recommendations