Bonjour a tous,
Ce n'est peut-etre pas le moment de lancer une discussion, mais je suis
curieux de savoir si vous avez deja teste l'impact de l'ordre des donnees
dans les matrices sur les resultats. Ceci m'a ete souffle par un message
recu de la liste Taxacom. Qu'en est-il d'ADE ?
Nicolas Bailly
____________________________________________________________
Museum National d'Histoire Naturelle,
Laboratoire d'Ichtyologie Generale et Appliquee
43, rue Cuvier, 75231 Paris Cedex 05, France
Tel: (33 1) 40 79 37 63 / 40 79 37 49 Fax: 40 79 37 71
Telex: MUSNAHN 202 641 F E-Mail: bailly@mnhn.fr
____________________________________________________________
Look at the Museum server http://www.mnhn.fr
On Tue, 6 Aug 1996, Utteridge Timothy Michael Arthur wrote:
> Dear Taxacomers,
>
> Having read Parnell & Waldren's paper about Detrended correspondence
> analysis (Taxon 45(1): 71-84. 1996) I was wondering if anyone has any
> experience of using DCA with taxonomic data. Parnell & Waldren list only
> advantages, there must be a few disadvantages or problems with using DCA.
> Comments on 'friendliness' of computer programs would be appreciated.
>
> Tim Utteridge
> Dept. of Ecology & Biodiversity
> University of Hong Kong
>
> Tel: +852 2857 9912
> Fax: +852 2559 5984
>
There are two opinions on DCA: the second axis arch should be flattened,
or you shouldn't use an algorithm that causes such non-independence
between axes. Wartenberg et al. 1987 (Am Nat 129:434-444) point to some
negatives, and see numerous references in Tausch et al. 1995 for other
detractors.
A more general issue is what do the results of any ordination mean, and
how much confidence can one place in the results? Ordination science does
not use inferential statistics, so true confidence intervals ae hard to
come by. So far as what the results mean, Tausch et al. (1995) J. Veg.
Sci 6:897-902 have shown that any software package that conducts
ordination analyses like PCA, DCA, and classification (cluster) analyses
such as UPGMA or TWINSPAN have a MAJOR problem: their results depend on
data entry order (which species comes first in the matrix, or which
character is character #1). That means that the software may settle on
a different answer depending on a completelty arbitrary characteristic of
the data. The problem is due to the fact that multidimensional analyses
are computationally complex, so the programmers have to use various
estimation steps for eigenvalues. If they didn't, the programs would take
much longer to give an answer.
So, before settling on any classification or ordination with your data, it
is wise to randomize it a couple of hundred times or so to see what kind
of diffferent results you'd get with different entry order. So much time
is spent collecting data of all types; we shouldn't produce or accept
potentially arbitrary results.
James
Tausch, R.J., D.A. Charlet, D.A. Weixelman, and D.C. Zamudio. 1995.
Patterns of ordination and classification instability resulting from
changes in input data order. J.Veg. Sci. 6:897-902.
Robin Tausch can be sent reprint requests at
USDA Forest Service
Intermountain Research Station
920 Valley Road
Reno, NV 89512 USA
_______________________________________________________________________________
\ / / \ / JAMES LYONS-WEILER ______________
\/ / \/ |..............|
\ / / |..............|
\/ / DOCTORAL PROGRAM IN |..............|
\ / ECOLOGY, EVOLUTION, AND |...***........|
\ / CONSERVATION BIOLOGY |..*****.......|
\ / |.******.......|
\/ 1000 VALLEY ROAD/186 |********......|
______________ THE UNIVERSITY OF --------------
| will perform | NEVADA, RENO
| statistical | RENO, NEVADA 89512-0013
| phylogenetic |
| analyses for | "(Biology) is not religion; if it were, we'd
| food | have a much easier time raising money."
-------------- -Leon Lederman
_______________________________________________________________________________
This archive was generated by hypermail 2b30 : Sat Feb 10 2001 - 10:21:20 MET