On Tue, 2007-05-22 at 00:35 +0000, dhinds at sonic.net
wrote:> I'd like to get a measure of goodness of fit for a heirarchical
> clustering result from hclust. Something that would indicate the
> extent to which the dendrogram accurately represents the original
> dissimilarity matrix. Is there an easy way to do this?
>
> Or, does anyone have code for computing distances between nodes given
> an hclust structure? So far, my searches have come up dry.
>
> -- David Hinds
Try ?cophenetic which calculates the cophenetic distances of a
hierarchical cluster analysis. The example on that help page shows how
to use the function to get the correlation between the original
distances and cophenetic distances.
HTH
G
--
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
Gavin Simpson [t] +44 (0)20 7679 0522
ECRC [f] +44 (0)20 7679 0565
UCL Department of Geography
Pearson Building [e] gavin.simpsonATNOSPAMucl.ac.uk
Gower Street
London, UK [w] http://www.ucl.ac.uk/~ucfagls/
WC1E 6BT [w] http://www.freshwaters.org.uk/
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%