Jan Baetens wrote: > Hi all, > > Currently, I'm using the built-in clustering algorithm of Mathematica 7, > though it isn't clear for me which algorithm it actually is since this > is not mentioned in the extended help pages. Presumably, it's the normal > K-means clustering but I'm not sure. > > As such, I'd like know whether someone knows which implementation is > used in Mathematica 7 for data clustering. > > Thanks, > > Jan > >
The default is k-medoids. Agglomerative clustering is also included as a method option. Brief discussion of the methods is included in the documentation. This can be found by entering
Here are some references about the methods that you might also find useful:
L. Kaufman and P. J. Rousseeuw, Finding Groups in Data: An Introduction to Cluster Analysis, New York: John Wiley & Sons, 1990.
P. J. Rousseeuw, Silhouettes: A Graphical Aid to the Interpretation and Validation of Cluster Analysis, J Comput. Appl. Math., 20, 1987, 5365.
R. Tibshirani, G. Walther, and T. Hastie, Estimating the Number of Clusters in a Dataset Via the Gap Statistic. Stanford Univ. Tech. report. March 2000. (published Journal of the Royal Statistical Society, B, 63, 2001, 411423.)