Full Width at Half Max (FWHM) of a peak in JMP

Question

FWHM was a common calculation I had to do in my former life on X-ray diffraction.  I had a customer ask if this is doable in JMP.  I know of a way that is not pretty where you find a maximum using summary statistics then you can get to the answer with a bit of hand calculation.  I saw this same question on the community without an answer I wanted to check to see if anyone know of a more elegant way to do this kind of calculation in JMP?

MikeD_Anderson · Answer

SASfit!  Yikes!  Haven't seen that for a while.  Some labmates used that.  I just used Fityk (http://fityk.nieto.pl).  I really hope that FDE will incorporate the functionality these other packages have.  It's so close already.

M

Peter_Hersh · Answer

Thanks Brady and Mike,

I think this will get him started.  I also agree with Mike that the Voigt fit would be great.

http://sasfit.ingobressler.net/manual/Voigt

brady_brady · Answer

Peter, something like this might work, depending on the shape of the curve, and if FWHM is as illustrated in the picture here:

https://en.wikipedia.org/wiki/Full_width_at_half_maximum

Names Default To Here(1);

//set up epsilon
e = 0.01;

//use a sample x and y
x = (0::6.28::.01)`;
y = cos(x)+1;

//find fwhm
hm = min(y) + (max(y)-min(y))/2;

xvals = x[loc(hm-e <= y <= hm+e)];

fwhm = max(xvals) - min(xvals);