**Next message:**Jaemin Park: "Re: MAJ98_Content"**Previous message:**Daniel Chessel: "MAJ98_Contenu"**Next in thread:**Jaemin Park: "Re: MAJ98_Content"**Reply:**Jaemin Park: "Re: MAJ98_Content"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ]

Hello, to our english speaking ADE-users

****************************************************************************

***********

The main modifications in the version 98 concern:

1) Correspondence analysis of correspondence with a column of reference, in

the CA family on models (Escofier, B. 1984. Analyse factorielle en

référence à un modèle.

Applications à l'analyse d'un tableau d'échanges. Revue de Statistique

Appliquée: 32, 4, 25-36). Implementation in COA: ColumnReference,

Detailed explanations in the form Thema2C.

2) PCA on fuzzy variables. The question is posed by the manipulation

of species-traits or species-ecological profiles tables. The option

FuzzyVar: Fuzzy Centring centers these tables in a logic of PCA. The

option PCA: Fuzzy PCA makes a PCA centered using assistance with the

specific interpretation. These options allow to process these tables by

way of K-sets methods. This operation is explained in the form Thema81

(Part 8 - under way development). The option Tables: FuzzyVariables

allows to visualize these tables.

3) Classic canonical analysis. It is available in Canonical: Canonical

correlations Analysis. The option reproduces the essential of calculations

described in the work of reference (Gittins, R. 1985. Canonical analysis, a

review with applications in ecology. Springer-Verlag, Berlin. 1-351)

4) Generalized canonical analysis, extension of the previous analysis for

K-tables (Carrol, J.D. 1968. A generalization of canonical correlation

analysis to three or more sets of variables. Proceeding of the 76th

Convention of the American

Psychological Association: 3, 227-228). Interpretation assistance

uses proposals of Casin, Ph. & Turlot, J.C. 1986. Une présentation de

l'analyse canonique généralisée dans l'espace des individus. Revue de

Statistique Appliquée: XXXV, 3, 65-75.

5) Matrices of correlation between tables. Canonical: CC coefficients

calculates matrices of correlation coefficients of the canonical

analysis, Canonical: RLS coefficients calculates matrices of coefficient

RLS (Kiers, H.A.L, Cléroux, R. & Ten Berge, M.F. (1994) Generalized

analysis based on optimizing matrix correlations and a relation with IDIOSCAL.

Computational Statistics and Data Analysis: 18, 331-340) and Canonical:

RV coefficients calculates matrices of RV coefficents. This last option is

a lightened version of the program ACT-STATIS centered on the

investigation of the interstructure. The module form gives definitions and

references.

Thank you to H. Kiers for its theoretical indications reported

in the module form.

6) Processing of lists of occurrences. The module OccurData allows to

approach ecological data when there are no longer samples, but solely

lists of occurrences (for example the recording of collection copies, the

plant collections,Š). OccurData: Read_Occur_File initiates procedures,

OccurData: CANOCO-Classes analyzes the list of occurrences according to a

partition (for example geographical), OccurData: CANOCO-Table analyzes the

list of occurrences when one has

environmental records associated to each occurrence, OccurData:

CANOCO-Subspace makes the same thing by using projeciton. The form Thema49

gives reasons for the choice of CANOCO as title of these analyses because

proposed analyses are true canonical analyses. This allows to indicate in

the form Thema4A when and why canonical correspondence analysis is a

canonical analysis. These innovations are part of the thesis of C.

Gimaret. OccurData: Array_to_Occur allows to transform the

species-by-samples tables in list of occurrences and to see that to choose

the CA and the CCA in ecology it is essentially considered that the

statistical unit is rather the occurrence of a taxon than the sample

itself. Such tools were formerly introduced by Green, R.H. 1971. A

multivariate statistical approach to the Hutchinsonian niche: bivalve

Molluscs of Central Canada. Ecology: 52, 543-556.

7) Analysis of the match between K tables and one table. The module

K+1 allows to introduce operators of co - inertia (X1, Y), (X2, Y),Š, (Xk,

Y) in an analysis of multiple coinertia analysis what finds the analysis of

concordances of Lafosse, R. & Hanafi, M. 1997. Concordance d'un tableau

avec K tableaux: définition de K+1 uples synthétiques. Revue de

Statistique Appliquée: (in press). With these implementations, ADE-4

contains now a panoply complete enough of K-sets methods.

Miscellaneous modifications:

CategVar: Read Categ File

files moda.cat and moda.123 are in option

COA: Internal COA

formats of publishing were improved

FilesUtil: Row_Duplicate

to copy rows of a table in function of a qualitative variable (tsee the

card of the pile info.hyp)

Graph1D: Match_2_var

to represent two numerical codes on one axis (example Thema49 p. 24)

Graph1D: Between_Graph

to represent a numerical code on one axis according to a partition (example

Thema49 p. 26)

Graph1D: Scal_2_scor

to represent two numerical codes on one axis (after standardization)

KTabUtil: MatchKTab-Tab

to associate K-tables and one table

Levels: Selected value

to trace a level curve given its value

Levels: Evenly spaced values

to trace level curves evenly spaced

OccurData: Rare_Species

to extract rare species from a list of occurences

Projectors: XY- > Orthonormal Basis - to get an orthonormal basis from

polynomials in x and y and to introduce the spatial coordinates in

canonical analyses (Wartenberg, D.E. 1985a. Canonical trend surface

analysis: a method for describing geographic

pattern. Systematic Zoology: 34 (3), 259-279)

Scatters: Labels

to trace a '+' sign in each point by typing value 3 in the dialogue box

'Draw points'

****************************************************************************

***********

HAVE FUN!

_____________________________________________

Sylvain Dolédec

Universite Claude Bernard Lyon 1

Bat 401C

69622 Villeurbanne CEDEX - France

Tel : (33) 4 72 43 13 63

Fax : (33) 4 72 43 11 41

_____________________________________________

ADE-4 package is available on the Internet

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

_____________________________________________

**Next message:**Jaemin Park: "Re: MAJ98_Content"**Previous message:**Daniel Chessel: "MAJ98_Contenu"**Next in thread:**Jaemin Park: "Re: MAJ98_Content"**Reply:**Jaemin Park: "Re: MAJ98_Content"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ]

*
This archive was generated by hypermail 2b30
: Sat Feb 10 2001 - 10:22:01 MET
*