Re: Problem avec les donnees pour ACM

From: Stephane Dray (
Date: Tue Apr 22 2003 - 16:21:34 MEST

>Dear Stephane,
>Thank you very much for your help. Excuse me first to write the message in
>English - I have not written message in French since 1992 and the keyboard
>here is not convenient neither. Actually, we are carrying out the
>so-called Constant Row Total-Multiple Correspondence Analysis (Guinand
>1996, Biological Journal of the Linnean Society, vol. 58, pp173) using the
>Fuzzy Correspondence Analysis program in the ADE-4 package. We use the
>TextToBin-BlocTo1Categ option to make a binary category file, and the
>input text file has only one column. We need to prepare a few files before
>we can use FuzzyVar-Read Fuzzy File to prepare the .fuz file for Fuzzy
>Correspondence Analysis. With small data sets, we don't have any problem.
>But when processing a large data set, truncated files were produced from
>'TextToBin-BlocTo1Categ' step. ADE-4 said that the binary matrix file
>(the raw data file) is too large to be read by ADEbin. The data have 26
>rows, and 8 variables with 35, 16, 46, 35, 26, 70, 64 and 27
>categories, respectively (319 columns in total). There should be no
>problem with computer power. I'm running ADE-4 on a Windows-XPP PC (DELL
>Optiplex, Pentium 4, CPU3.06GHz, RAM 1.00Gb)
>Could you, and anyone else on the list, help me on this issue please?
>Thank you very much..

I think that there is no problem with your data... problem comes from
ADEBin. I understand that TextToBin-BlocTo1Categ option works but the
error message appears when you want to see the file .. this is
"normal".. Indeed, ADEBin is a module to view small binary files. Try
to convert your binary file into a text file with ADETrans or
BintoText. Is the text file truncated ?? I hope that the answer is
no. Another way to see if there is a problem: If you use VerifCateg,
what ADE-4 tells you ?
If you prefer, ade4 is now a package of R statistical software and
fuzzy correspondence analysis is available by the function dudi.fca


Stéphane DRAY
Biométrie et Biologie évolutive - Equipe "Écologie Statistique"
Universite Lyon 1 - Bat 711 - 69622 Villeurbanne CEDEX - France

Tel : 04 72 43 27 56 Fax : 04 78 89 27 19 04 72 43 27 57 E-mail : --------------------------------------------------------------- Web ---------------------------------------------------------------

This archive was generated by hypermail 2b30 : Tue Sep 07 2004 - 13:45:20 MEST