Closed mducle closed 4 months ago
Discussed - @mducle to work on mex then @RichardWaiteSTFC to review and merge. In a separate PR future plan is to:
fitpow
to be a separate class not method of spinw (and put in swfiles) - input struct will be setting attributes in that classClosed in favour of #174
Adds a method to fit powder data.
To test, run this script with the attached data: mnf2.mat.zip
cutpow.m
:It should take 1-2min to run the fit.
Note that for some reason,
lm
andlm2
(there are 3 implementations of the Levenberg-Marquardt algorithm inndbase
) doesn't work butlm3
does work.