hello, I would like to compare indexes in order to see which clustering algorithm (k- means, k-medoid, hierarchical clustering with euclidean and pearson distance) is more efficient. So I want to calculate the c-index and the db in all those cases(library cclust). To do that I use the cclust function. How can I run this function with pearson distance for k-means? How can I generalize cclust for hierarchical clustering and k-medoid? The class cclust object has a list of 13 components. Can anybody tell me how these components are calculated? thanks for everything. Emmanuelle Anthoine. ---------------------------------------------------------------------- ---------- Université de Bretagne sud http://www .univ-ubs.fr/