**Next message:**umwelt.institut.hoexter: "Clusters - Prepare convex hulls"**Previous message:**L.Blanc: "Re : ACOM"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ]

<x-charset ISO-8859-1>Hi to ADE users,

I wonder whether the question of Bruno Rossaro was answered.

Anyway, concerning the operation for getting a trait-by-sites matrix

(proportion of q column-modalities of r traits in n row-sites) from a

traits-by-species matrix (proportion of q column-modalities of r traits of t

row-species) and a sites-by-species matrix (Abundance of t row-species in n

column-sites) at least two possibilities are available in ADE-4.

The first one is to use the the FuzFCMean option of the Fuzzyvar module.

This requires first to process the trait-by-species matrix using the Read

Fuzzy File option of the same module and to perform a correspondence

analysis on the sites-by-species matrix (file .fcta required by the

FuzFCMean option). Note than if you make the co-inertia analysis between the

traits-by-species matrix and the sites-by-species matrix matrix, you have a

way for testing the significance of the trait-by-sites matrix (see Doc

module of ADE-4).

If you do not want to perform a correspondence analysis on the

sites-by-species matrix, you may compute the product of the two matrices

more straight. This just requires that you process the trait-by-species

matrix using the Read Fuzzy File option of the Fuzzyvar module and you

transform the sites-by-species matrix into frequencies (by column option).

You can then multiply the two matrices using the Diagonal Inner Product of

the MatAlg module (with Inner product D = 1,1,...,1).

The traits-by-species matrix can then be processed by Fuzzy FCA, Fuzzy PCA

or else (after being processed by the reading option).

All the very best,

Sylvain Dolédec

----------------------------------------

UCB Lyon 1

CNRS - Ecologie des Hydrosystèmes Fluviaux

Bât 401C

43 Bd du 11 novembre 1918

69622 Villeurbanne

phone: +33 (0) 472431363

fax: +33 (0) 472431141

----------------------------------------

http://ecolstat.univ-lyon1.fr/equipe/accueil/

http://pbil.univ-lyon1.fr/ADE-4/ADE-4.html

----------

*>De : Rossaro Bruno <rossaro@mailserver.unimi.it>
*

*>À : adelist@biomserv.univ-lyon1.fr
*

*>Objet : fuzzy corresp analysis
*

*>Date : Mar 23 mai 2000 16:12
*

*>
*

*> Dear collegues,
*

*> I have read with interest the papers published in Freshwat. Biol.
*

*> using fuzzy correspondence analysis.
*

*> charvet@lyon.cemagref.fr
*

*> In particular:
*

*> Authors
*

*> Charvet S. Statzner B. Usseglio-Polatera P. Dumont B.
*

*> Author e-mail Address
*

*> charvet@lyon.cemagref.fr
*

*> Title
*

*> Traits of benthic macroinvertebrates in semi-natural French streams: an
*

*> initial application to biomonitoring in Europe
*

*> Source
*

*> Freshwater Biology. 43(2):277-296, 2000 Feb.
*

*> In this paper I have not well understood how fuzzy correspondence analysis
*

*> was carried out on a trait category matrix
*

*> weighted on species abundances.
*

*> Suppose you have a trait matrix with t species and m modalities (tMm) and
*

*> an abundance matrix with n sites and t species
*

*> (nYt). It seems that something such as a product matrix nYtMm was
*

*> performed, and fuzzy correspoindence analysis was carried out on this matrix.
*

*> Can somebody help me ?
*

*> Sincerely yours
*

*> Bruno Rossaro
*

*> _______________________________________
*

*> prof. Bruno Rossaro
*

*> Department of Biology Section Ecology
*

*> Univ. of Milano
*

*> via Celoria 26 I 20133 Milano (office)
*

*> v.le Murillo 21 I 20149 Milano (private address)
*

*> tel 039_2_2_26604_371
*

*> fax 0039_2_26604_361
*

*> e-mail: Bruno.Rossaro@unimi.it
*

*> http://www.sezionecologia.unimi.it
*

*>
*

</x-charset>

**Next message:**umwelt.institut.hoexter: "Clusters - Prepare convex hulls"**Previous message:**L.Blanc: "Re : ACOM"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ]

*
This archive was generated by hypermail 2b30
: Mon Feb 12 2001 - 09:24:55 MET
*