[Ifeffit] Re: PCA

[Webb, Samuel M.]

Hi Ethan,

Glad to hear you've been checking out the PCA routines in SixPACK.  I'm away from the office until after the July 4th holiday, so email access will be a little impeded.  

Your're right, the "save component" menu feature doesn't work, much to my surprise.  I'll have this fixed in my next release.  However, until that time (I won't be able to make an improved version until post July 5th), here is a quick work-around.  You can plot the components, then use the "copy plot data to clipboard" option in the plot menu.  This will place all current plots on the clipboard, which can then be pasted into any other application you'd like for further examination.  If you want to save the components for further use in Sixpack or Athena/Artemis, any 2 column ascii format will work.

As for the target transform, I can put that information of that fit into the next relase as well.  As a temporary fix again, you can take the saved components (saving each of the plots above into its own text file) and then do a linear combo fit to get that data, using the PCA components as the components and the target transform file as the sample.  This should give you the same fit as the target transform does in the PCA routine.

Thanks for the feedback.  Its always good to get comments from other users -- particularly on things that I didn't think of on the first time around.  I'll try to have those improvements in the PCA released when I get back to my office.

Sam

********************************************
Sam Webb
SSRL/SLAC Bldg. 137, MS 69
2575 Sand Hill Rd.
Menlo Park, CA 94025

Phone:  (650)-926-3734
FAX:    (650)-926-4100
E-mail: samwebb at slac.stanford.edu
*******************************************