Using fit data printed in history

Share icon

Mon, 01/06/2014 - 04:51 pm

I used the Gauss2D feature to fit lines onto an image graph, and the following information is printed into the history:

•CurveFit/NTHR=0 Gauss2D temperaturedata0 /D
Fit converged properly
--Curve fit with constraints--
No constraints active or violated
fit_temperaturedata0= W_coef[0]+W_coef[1]*exp((-1/(2*(1-W_coef[6]^2)))*(((x-W_coef[2])/W_coef[3])^2 + ((y-W_coef[4])/W_coef[5])^2 - (2*W_coef[6]*((y-W_coef[4])*(x-W_coef[2]))/(W_coef[3]*W_coef[5]))))
W_coef={148.16,201.92,374.55,18.493,285.51,19.406,0.54951}
V_chisq= 5.5801e+008;V_npnts= 307200;V_numNaNs= 0;V_numINFs= 0;
V_startRow= 0;V_endRow= 639;V_startCol= 0;V_endCol= 479;
W_sigma={0.0779,1.96,0.18,0.18,0.189,0.189,0.0096}
Coefficient values ± one standard deviation
z0 =148.16 ± 0.0779
A =201.92 ± 1.96
x0 =374.55 ± 0.18
xWidth =18.493 ± 0.18
y0 =285.51 ± 0.189
yWidth =19.406 ± 0.189
cor =0.54951 ± 0.0096


Is there a way I can use these outputs, like xWidth and yWidth? Can they be saved as variables somewhere? I'm trying to make a function that plots several widths of 2D graphs into a new graph, to see how a cloud is shrinking in time.
Yes, but not quite like you think. Rather than using the print output you can utilize the coefficient waves. Look at your data folder after the fit. there are two waves W_coef and W_sigma which hold all the information you need. Look closely at your print output and you will notice that W_coef holds all the fit values and W_sigma contains the respective deviations. To go by your example, to extract xWidth you just need to save the fourth value of W_coef (and maybe W_sigma if needed) somewhere before doing the next fit.

January 6, 2014 at 09:09 pm - Permalink