RGARONO@kentvm.bitnet ("Ralph J. Garono") (06/13/89)
Decorana is a fortran program for detrended correspondence analysis and
reciprocal averaging written by M.O. Hill of Cornell University. The
software and documentation are available from Cornell.
I am presently involved in using decorana to evaluate plant community class-
ification here in Ohio. I have also used this multivariate technique to
classify caddisfly fauna of ohio peatlands.
---Ralph J. GaronoD.NYE@qmc.ac.uk (06/13/89)
Could anyone advise on the origin and /or availability of two packages called TWINSPAN and DECARANA which I understand to be associated with cluster analysis. Any other information would also be appreciated e.g. hardware requirements. My e-mail address is: D.NYE@QMC.AC.UK Cheers, David Nye, Queen Mary College Computer Centre, London.
LEBOUL@buclln11.bitnet (Eric Le Boulenge) (06/13/89)
More information on software for cluster analysis: there are two other software performing numerous types of cluster and other multivariate analyses, you might consider useful. Here are the references: 1. Package "CANOCO" by Cajo ter Braak and colleagues at Wageningen (the Netherlands): versions for MacIntosh, IBM-PC and mainframes available, contact Cajo at the following EARN (Bitnet) address: STATLB at HWALHW50 Mainly Principal Components, Correspondence Analysis (detrended, constrained upon request). 2. "R" Package by Pierre Legendre and Alain Vaudor at Montreal University, (Canada): versions for MacIntosh and for mainframe available, contact Pierre at the following BITNET address: 3771 at UMTLVR.BITNET Many clustering algorithms based on many distance/similarity measures, diverse association tests (Mantel type), correlograms, periodograms, and data management routines. I know for sure the second package is distributed free of charge; there is some payment requested for the first one. Sorry to talk in a distribution list of a package that could possibly be considered as of "commercial" type: I do not intend to advertise for these, merely to inform people interested in cluster analysis, of the existence of software that may help them. And I have no commitment at all with the authors of these software... We just use these packages in our lab.... Sincerely, Eric Le Boulenge. Unite d'Ecologie, Universite Catholique de Louvain, Louvain-la-Neuve, Belgium.