Dear list member!
As I am new to this field I have some rather general questions regarding
ADE especially on PLS (Partial Least Squares).
It seems like PLS is most commonly used in chemistry (analytical/organic),
but do anyone know if it is used in biology (especially in environmental
monitoring and/or aquatic siences). Most of the articles that I have found
are either written by chemists or by statisticians (methodological
articles), but very few by biologist. WHY IS IT SO? Is PLS only a marginal
method with very limited use? According to what is written about it, it
seems rather to be a tremendously useful tool, almost to good to be true.
Which differences are there between ADE and SIMCA (esp. concerning PLS),
except that SIMCA only runs on PC´s (SIMCA seems to me to be the most
commonly used software for this kind of analyses in chemistry and ADE
mostly used by french speaking ecologist, is that true?). The only
difference as I have understood so far is how they handle missing values
(ADE writes zeros or -999 when the data is imported). In PCA/PCR I have
also found the minor difference (at least when n is large) that ADE uses
1/n as weight (as default), while other softwares like SAS and JMP uses
1/df as default.
When I try to make a permutation test in PLS2 the program always abruptly
terminates giving me a message that there has been an error 2 (i.e. an
address error). Also when I perform the modelling within PLS2 the program
aborts, but luckily after the analysis is done so the results can be found
in the appropriate files. I am using the PowerMac version of PLS2 on my PM
7200/90 with 120 MB RAM and I have tried to down-load new copies of PLS2
with the same result. Do anybody else have the same problem?
I am looking forward to your answers.
Lars Sonesten MSc
Institute of Limnology, Uppsala University
Norbyvaegen 20, SE-752 36 Uppsala, Sweden
Tel: int.code +46 - 18 - 471 27 28
Fax: int.code +46 - 18 - 53 11 34
Internet site: <http://www.limno.uu.se/> -- Institute of Limnology
This archive was generated by hypermail 2b30 : Sat Feb 10 2001 - 10:22:00 MET