Clustering for Microanalytical Data
Clustering is a chemometrical technique, used for the grouping or classification of objects characterized by multivariate analytical data. The hierarchical and non-hierarchical algorithms are discussed. Among the hierarchical methods the average linkageand Ward’s method are preferred, while the preferred non-hierarchical method is MASLOC. The latter is the only method with a build-in rule for the selection of significant clusters. The importance of combining clustering with other multivariate techniques, such as factor analysis, is stressed.
KeywordsHierarchical Cluster Significant Cluster Seed Point Iron Meteorite Natural Cluster
Unable to display preview. Download preview PDF.
- 3.D.L. Massart and H. De Clercq, Anal. Chem. 46, 1988, 1974.Google Scholar
- 4.D.L. Massart, A. Dijkstra and L. Kaufman. Evaluation and Optimization of Laboratory Methods and Analytical Procedures, Elsevier, Amsterdam 1978.Google Scholar
- 6.E.W. Forgy, cited by M.R. Anderberg, Cluster Analysis for Applications, Academic Press, New York, 1973.Google Scholar
- 7.J.B. McQueen, cited by M.R. Anderberg, see ref. 6Google Scholar
- 8.Clustan, developed by D. Wishart and obtainable from The Clustan Project, Computer Centre, University College of London, 19 Gordon Street, London WC1HOAHGoogle Scholar
- 9.G.H. Ball and D.J. Hall, Isodata, A novel method of Data Analysis and Pattern Classification, AD 699616, Stanford Res. Inst., Menlo Park, California (1965).Google Scholar
- 10.D.L. Massart, L. Kaufman and D. Coomans, Anal. Chim. Acta, in press.Google Scholar
- 12.A.M. Massart-Leen and D.L. Massart, to be published.Google Scholar
- 13.H.K. Esbensen and Y.F. Buchwald, to be published.Google Scholar
- 14.S. Wold, Chapters 19 and 20, of ref. 4.Google Scholar