APL-STAT, A Do-It-Yourself Guide to Computational Statistics Using APL
Martin Schatzoff · Technometrics · 1982
methods by means of artificially generated data. The book features a variety of methods. It is regrettable that Everitt has dropped the group average method (UPGMA) from the new edition, since it is the single most commonly applied method in the biological sciences. The centroid method, which he features instead, suffers from the disadvantage of resulting in reversals during hierarchic cluster analysis. The violation of the ultrametric can be seen in the graphic illustrations that are furnished in Everitt's book, but nowhere are the implications of this deficiency of the centroid method pointed out. Beginners will not find this book as useful as Spaith's Cluster-Analyse-Algorithmen, since it lacks a variety of illustrative data, the sample problems are simpler than they could have been, and there are no simple FORTRAN programs for clustering procedures.