Talk uploaded / rising interest in NNMA
Wellenreuther, Gerd
just wanted to tell you that I updated the MAHID page with a link to my
talk about Non-Negative Matrix Approximation (NNMA) given at the
European Conference on X-Ray Spectrometry (EXRS) in Portugal this June.
Apart from the daphni magna data it also shows data from a geological
sample containing monazite, which factors really beautifully. The direct
link is http://www.embl-hamburg.de/%7Egwellenr/NNMA%20EXRS%202010.ppt .
After that talk I was approached by several people interested in the
technique. In addition, Geert Silversmit, Andy Goldschmidt and I are
planning to work together on finding some kind of general protocol for
NNMA especially suited for fluorescence data. If anybody else wants to
join or just discuss his work ...
Cheers, Gerd
--
Dr. Gerd Wellenreuther
beamline scientist P06 "Hard X-Ray Micro/Nano-Probe"
Petra III project
HASYLAB at DESY
Notkestr. 85
22603 Hamburg
Tel.: + 49 40 8998 5701
Pete Jemian
Berd:
Thanks for the link to your PowerPoint presentation. I wonder about
another non-negative approach (actually a couple but let's not talk
about MaxEnt here).
TNNLS:
Totally nonnegative least-squares problems
- interior-point gradient methods
http://www.springerlink.com/content/n31j0625l0nth256/
I'm curious and a bit naive about XRF data reduction and image extraction.
Have you considered TNNLS (or some other method such as a Tikhonov
regularization) as a possibility and perhaps rejected it for some
reasons? It's used by some scientists, most of all I'm familiar with
its use in small-angle scattering data analysis.
Thanks,
Pete