Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If You Use a Screen Reader

This content is available through Read Online (Free) program, which relies on page scans. Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.

Model-Based Gaussian and Non-Gaussian Clustering

Jeffrey D. Banfield and Adrian E. Raftery
Biometrics
Vol. 49, No. 3 (Sep., 1993), pp. 803-821
DOI: 10.2307/2532201
Stable URL: http://www.jstor.org/stable/2532201
Page Count: 19
  • Read Online (Free)
  • Download ($14.00)
  • Subscribe ($19.50)
  • Cite this Item
Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
Model-Based Gaussian and Non-Gaussian Clustering
Preview not available

Abstract

The classification maximum likelihood approach is sufficiently general to encompass many current clustering algorithms, including those based on the sum of squares criterion and on the criterion of Friedman and Rubin (1967, Journal of the American Statistical Association 62, 1159-1178). However, as currently implemented, it does not allow the specification of which features (orientation, size, and shape) are to be common to all clusters and which may differ between clusters. Also, it is restricted to Gaussian distributions and it does not allow for noise. We propose ways of overcoming these limitations. A reparameterization of the covariance matrix allows us to specify that some, but not all, features be the same for all clusters. A practical framework for non-Gaussian clustering is outlined, and a means of incorporating noise in the form of a Poisson process is described. An approximate Bayesian method for choosing the number of clusters is given. The performance of the proposed methods is studied by simulation, with encouraging results. The methods are applied to the analysis of a data set arising in the study of diabetes, and the results seem better than those of previous analyses. A magnetic resonance image (MRI) of the brain is also analyzed, and the methods appear successful in extracting the main features of anatomical interest. The methods described here have been implemented in both Fortran and S-PLUS versions, and the software is freely available through StatLib.

Page Thumbnails

  • Thumbnail: Page 
803
    803
  • Thumbnail: Page 
804
    804
  • Thumbnail: Page 
805
    805
  • Thumbnail: Page 
806
    806
  • Thumbnail: Page 
807
    807
  • Thumbnail: Page 
808
    808
  • Thumbnail: Page 
809
    809
  • Thumbnail: Page 
810
    810
  • Thumbnail: Page 
811
    811
  • Thumbnail: Page 
812
    812
  • Thumbnail: Page 
813
    813
  • Thumbnail: Page 
814
    814
  • Thumbnail: Page 
815
    815
  • Thumbnail: Page 
816
    816
  • Thumbnail: Page 
817
    817
  • Thumbnail: Page 
818
    818
  • Thumbnail: Page 
819
    819
  • Thumbnail: Page 
820
    820
  • Thumbnail: Page 
821
    821