Psychometrika

, Volume 50, Issue 2, pp 159–179 | Cite as

An examination of procedures for determining the number of clusters in a data set

  • Glenn W. Milligan
  • Martha C. Cooper
Article

Abstract

A Monte Carlo evaluation of 30 procedures for determining the number of clusters was conducted on artificial data sets which contained either 2, 3, 4, or 5 distinct nonoverlapping clusters. To provide a variety of clustering solutions, the data sets were analyzed by four hierarchical clustering methods. External criterion measures indicated excellent recovery of the true cluster structure by the methods at the correct hierarchy level. Thus, the clustering present in the data was quite strong. The simulation results for the stopping rules revealed a wide range in their ability to determine the correct number of clusters in the data. Several procedures worked fairly well, whereas others performed rather poorly. Thus, the latter group of rules would appear to have little validity, particularly for data sets containing distinct clusters. Applied researchers are urged to select one or more of the better criteria. However, users are cautioned that the performance of some of the criteria may be data dependent.

Key words

classification stopping rules numerical taxonomy 

References

  1. Andrews, D. F. (1972). Plots of high-dimensional data.Biometrics, 28, 125–136.Google Scholar
  2. Arnold, S. J. (1979). A test for clusters.Journal of Marketing Research, 19, 545–551.Google Scholar
  3. Baker, F. B., & Hubert, L. J. (1975). Measuring the power of hierarchical cluster analysis.Journal of the American Statistical Association, 70, 31–38.Google Scholar
  4. Ball, G. H., & Hall, D. J. (1965).ISODATA, A novel method of data analysis and pattern classification. Menlo Park: Stanford Research Institute. (NTIS No. AD 699616).Google Scholar
  5. Beale, E. M. L. (1969).Cluster analysis. London: Scientific Control Systems.Google Scholar
  6. Binder, D. A. (1978). Bayesian cluster analysis.Biometrika, 65, 31–38.Google Scholar
  7. Blashfield, R. K., & Morey, L. C. (1980). A comparison of four clustering methods using MMPI Monte Carlo data.Applied Psychological Measurement, 4, 57–64.Google Scholar
  8. Bock, H. H. (1977). On tests concerning the existence of a classification. InFirst international symposium on data analysis and informatics (Vol. 2, pp. 449–464). Rocquencourt, France: IRIA.Google Scholar
  9. Calinski, R. B., & Harabasz, J. (1974). A dendrite method for cluster analysis.Communications in Statistics, 3, 1–27.Google Scholar
  10. Cohen, A. C. (1967). Estimation in mixtures of two normal distributions.Technometrics, 9, 15–28.Google Scholar
  11. Davies, D. L., & Bouldin, D. W. (1979). A cluster separation measure.IEEE Transactions on Pattern Analysis and Machine Intelligence, 1, 224–227.Google Scholar
  12. Day, N. E. (1969). Estimating the components of a mixture of normal distributions.Biometrika, 56, 463–474.Google Scholar
  13. Dubes, R., & Jain, A. K. (1979). Validity studies in clustering methodologies.Pattern Recognition, 11, 235–254.Google Scholar
  14. Duda, R. O., & Hart, P. E. (1973).Pattern classification and scene analysis. New York: Wiley.Google Scholar
  15. Edwards, A. W. F., & Cavalli-Sforza, L. (1965). A method for cluster analysis.Biometrics, 21, 362–375.Google Scholar
  16. Englemann, L., & Hartigan, J. A. (1969). Percentage points of a test for clusters.Journal of the American Statistical Association, 64, 1647–1648.Google Scholar
  17. Everitt, B. S. (1979). Unresolved problems in cluster analysis.Biometrics, 35, 169–181.Google Scholar
  18. Everitt, B. S. (1981). A Monte Carlo investigation in the likelihood ratio test for the number of components in a mixture of normal distributions.Multivariate Behavioral Research, 16, 171–180.Google Scholar
  19. Fleiss, J. L., Lawlor, W., Platman, S. R., & Fieve, R. R. (1971). On the use of inverted factor analysis for generating typologies.Journal of Abnormal Psychology, 77, 127–132.Google Scholar
  20. Fleiss, J. L., & Zubin, J. (1969). On the methods and theory of clustering.Multivariate Behavioral Research, 4, 235–250.Google Scholar
  21. Friedman, H. P., & Rubin, J. (1967). On some invariant criteria for grouping data.Journal of the American Statistical Association, 62, 1159–1178.Google Scholar
  22. Frey, T., & Van Groenewoud, H. (1972). A cluster analysis of the D-squared matrix of white spruce stands in Saskatchewan based on the maximum-minimum principle.Journal of Ecology, 60, 873–886.Google Scholar
  23. Fukunaga, K., & Koontz, W. L. G. (1970). A criterion and an algorithm for grouping data.IEEE Transactions on Computers, C-19, 917–923.Google Scholar
  24. Gengerelli, J. A. (1963). A method for detecting subgroups in a population and specifying their membership list.Journal of Psychology, 5, 457–468.Google Scholar
  25. Gnanadesikan, R., Kettenring, J. R., & Landwehr, J. M. (1977). Interpreting and assessing the results of cluster analyses.Bulletin of the International Statistical Institute, 47, 451–463.Google Scholar
  26. Good, I. J. (1982). An index of separateness of clusters and a permutation test for its statistical significance.Journal of Statistical Computing and Simulation, 15, 81–84.Google Scholar
  27. Goodall, D. W. (1966). Hypothesis testing in classification.Nature, 221, 329–330.Google Scholar
  28. Gower, J. C. (1975). Goodness-of-fit criteria for classification and other patterned structures. In G. Estabrook, (Ed.),Proceedings of the 8th international conference on numerical taxonomy. San Francisco: Freeman.Google Scholar
  29. Gower, J. C. (1981, June).Is classification statistical? Paper presented at the meeting of the Classification Society, Toronto.Google Scholar
  30. Hall, D. J., Duda, R. O., Huffman, D. A., & Wolf, E. E. (1973).Development of new pattern recognition methods. Los Angeles: Aerospace Research Laboratories. (NTIS No. AD 7726141).Google Scholar
  31. Hansen, R. A., & Milligan, G. W. (1981). Objective assessment of cluster analysis output: Theoretical considerations and empirical findings.Proceedings of the American Institute for Decision Sciences, 314–316.Google Scholar
  32. Hartigan, J. A. (1975).Clustering algorithms. New York: Wiley.Google Scholar
  33. Hartigan, J. A. (1977). Distribution problems in clustering. In J. Van Ryzin (Ed.),Classification and clustering. New York: Academic Press.Google Scholar
  34. Hartigan, J. A. (1978). Asymptotic distributions for clustering criteria.Annals of Statistics, 6, 117–131.Google Scholar
  35. Hill, R. S. (1980). A stopping rule for partitioning dendrograms.Botanical Gazette, 141, 321–324.Google Scholar
  36. Hubert, L. J., & Baker, F. B. (1977). The comparison and fitting of given classification schemes.Journal of Mathematical Psychology, 16, 233–253.Google Scholar
  37. Hubert, L. J., & Levin, J. R. (1976). A general statistical framework for assessing categorical clustering in free recall.Psychological Bulletin, 83, 1072–1080.Google Scholar
  38. Jain, A. K., & Waller, W. G. (1978). On the number of features in the classification of multivariate gaussian data.Pattern Recognition, 10, 365–374.Google Scholar
  39. Jancey, R. C. (1966). Multidimensional group analysis.Australian Journal of Botany, 14, 127–130.Google Scholar
  40. Johnson, S. C. (1967). Hierarchical clustering schemes.Psychometrika, 32, 241–254.Google Scholar
  41. Lee, K. L. (1979). Multivariate tests for clusters.Journal of the American Statistical Association, 74, 708–714.Google Scholar
  42. Lingoes, J. C., & Cooper, T. (1971). PEP-I: A FORTRAN IV (G) program for Guttman-Lingoes nonmetric probability clustering.Behaviorial Science, 16, 259–261.Google Scholar
  43. Marriot, F. H. C. (1971). Practical problems in a method of cluster analysis.Biometrics, 27, 501–514.Google Scholar
  44. McClain, J. O., & Rao, V. R. (1975), CLUSTISZ: A program to test for the quality of clustering of a set of objects.Journal of Marketing Research, 12, 456–460.Google Scholar
  45. Milligan, G. W. (1980). An examination of the effect of six types of error perturbation on fifteen clustering algorithms.Psychometrika, 45, 325–342.Google Scholar
  46. Milligan, G. W. (1981a). A Monte Carlo study of thirty internal criterion measures for cluster analysis.Psychometrika, 46, 187–199.Google Scholar
  47. Milligan, G. W. (1981b). A review of Monte Carlo tests of cluster analysis.Multivariate Behavioral Research, 16, 379–407.Google Scholar
  48. Milligan, G. W. (1981c, June).A discussion of procedures for determining the number of clusters in a data set. Paper presented at the meeting of the Classification Society, Toronto.Google Scholar
  49. Milligan, G. W. (1983). Characteristics of four external criterion measures. In J. Felsenstein, (Ed.),Proceedings of the 1982 NATO Advanced Studies Institute on Numerical Taxonomy (pp. 167–173). New York: Springer-Verlag.Google Scholar
  50. Milligan, G. W., & Sokol, L. M. (1980). A two-stage clustering algorithm with robust recovery characteristics.Educational and Psychological Measurement, 40, 755–759.Google Scholar
  51. Milligan, G. W., Soon, S. C., & Sokol, L. M. (1983). The effect of cluster size, dimensionality, and the number of clusters on recovery of true cluster structure.IEEE Transactions on Pattern Analysis and Machine Intelligence, 5, 40–47.Google Scholar
  52. Mojena, R. (1977). Hierarchical grouping methods and stopping rules: An evaluation.The Computer Journal, 20, 359–363.Google Scholar
  53. Morey, L., & Agresti, A. (1984). The measurement of classification agreement: An adjustment to the Rand statistic for chance agreement.Educational and Psychological Measurement, 44, 33–37.Google Scholar
  54. Mountford, M. D. (1970). A test for the difference between clusters. In G. P. Patil, E. C. Pielou, & W. E. Waters (Eds.),Statistical Ecology (Vol. 3, pp. 237–257). University Park, Pa.: Pennsylvania State University Press.Google Scholar
  55. Naus, J. I. (1966), A power comparison of two tests of non-random clustering.Technometrics, 8, 493–517.Google Scholar
  56. Orloci, L. (1967). An agglomerative method for classification of plant communities.Journal of Ecology, 55, 193–206.Google Scholar
  57. Perruchet, C. (1983).Les épreuves de classifiabilité en analyses des données [Statistical tests of classificability] (Tech. Rep. NT/PAA/ATR/MTI/810). Issy-Les-Moulineaux, France: C.N.E.T.Google Scholar
  58. Ray, A. A. (Ed.). (1982).SAS user's guide: Statistics. Cary, North Carolina: SAS Institute.Google Scholar
  59. Ratkowsky, D. A., & Lance, G. N. (1978). A criterion for determining the number of groups in a classification.Australian Computer Journal, 10, 115–117.Google Scholar
  60. Rohlf, F. J. (1974). Methods of comparing classifications.Annual Review of Ecology and Systematics, 5, 101–113.Google Scholar
  61. Rubin, J. (1967). Optimal classification into groups: An approach for solving the taxonomy problem.Journal of Theoretical Biology, 15, 103–144.Google Scholar
  62. Sarle, W. S. (1983).Cubic clustering criterion (Tech. Rep. A-108). Cary, N.C.: SAS Institute.Google Scholar
  63. Scott, A. J., & Symons, M. J. (1971). Clustering methods based on likelihood ratio criteria.Biometrics, 27, 387–397.Google Scholar
  64. Sneath, P. H. A. (1977). A method for testing the distinctness of clusters: A test of the disjunction of two clusters in Euclidean space as measured by their overlap.Mathematical Geology, 9, 123–143.Google Scholar
  65. Sneath, P. H. A., & Sokal, R. R. (1973).Numerical taxonomy. San Francisco: Freeman.Google Scholar
  66. Sokal, R. R., & Sneath, P. H. A. (1963).Principles of numerical taxonomy. San Francisco: Freeman.Google Scholar
  67. Thorndike, R. L. (1953). Who belongs in a family?Psychometrika, 18, 267–276.Google Scholar
  68. Wolfe, J. H. (1970). Pattern clustering by multivariate mixture analysis.Multivariate Behavioral Research, 5, 329–350.Google Scholar
  69. Wong, M. A. (1982). A hybrid clustering method for identifying high-density clusters.Journal of the American Statistical Association, 77, 841–847.Google Scholar
  70. Wong, M. A., & Schaak, C. (1982). Using the Kth nearest neighbor clustering procedure to determine the number of subpopulations.Proceedings of the Statistical Computing Section, American Statistical Association, 40–48.Google Scholar

Copyright information

© The Psychometric Society 1985

Authors and Affiliations

  • Glenn W. Milligan
    • 1
  • Martha C. Cooper
    • 1
  1. 1.Faculty of Management SciencesThe Ohio State UniversityColumbus

Personalised recommendations