Comparison of Classifications with the Data from which they are Derived
The application of clustering methods to ecological problems, which take place in a continuum, is sometimes puzzling because of lack of statistical methods for validating and comparing results. The aim of this paper is to propose an empirical way to compare the results from different hierarchical classifications with the original matrix of proximities. This method should permit us to evaluate the modifications of the initial structure and, finally, to understand the specificity of clustering methods.
Unable to display preview. Download preview PDF.
- Cailliez, F., and J.P Pages. 1976. Introduction A l’analyse des données. Société de mathématiques appliquées et de Sciences humaines. Paris XXI 616 p.Google Scholar
- Gower, J.C. 1966. Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53: 325–338.Google Scholar
- Sneath, P.H.A. 1974. Phylogeny of micro-organisms. Symposia of the society for general Microbiology. N XXIV. Evolution in the microbial world: 1–39.Google Scholar