[Ifeffit] SixPack PCA

Andrew acampo2 at tigers.lsu.edu
Thu Apr 29 14:48:02 CDT 2010


I'm by no means a statistical expert. There are some rules of thumbs in the
literature regarding what Spoil value to believe in or not. 

Some resources that you can look up:

1. Beauchemin, S.; Hesterberg, D.; Beauchemin, M., Soil Sci Soc Am J 2002,
66 (1), 83-91.
(( in the above work they included standards which had SPOIL values greater
than 6, so they would have rejected a standard with a value lower than 6))
2. Malinowski, E. R., Factor Analysis in Chemistry. Wiley: New York, 2002; p

I might be wrong on this but when these programs use the formulas for
calculating F-values and SPOIL values, they assume that all points are
independent. One question that comes up sometimes is "How many independent
points are there in a XANES spectrum?" In EXAFS many people use the Nyquist
criterion. And what I'm getting at is you have to be cautious in using a
rule of thumb for the XANES analysis. If there are other techniques you can
use to help you identify what is in your sample try those as well. 

The manual on Athena also points out that the number of independent points
collected in a XANES spectrum is difficult to determine. Without additional
information: the quality of the data collected (the signal-to-noise), what
edge you collected your spectra, are you fitting standards that you
collected in the same set of runs as your samples, I think it's difficult
for anyone to go into further depth. 

Please look up the above two references, this should at least give you a
better idea.  

Hope this helps!
Andrew Campos

More information about the Ifeffit mailing list