Performing a Curve Fit
Once you have completed your estimates for the curve fit bands, you can
optimize the fit by clicking the Fit button in the Select Bands dialog. This
instructs the computer to use non-linear regression algorithms in order to “fine tune” the parameters of each band in such a way that the sum of the bands matches
the experimental data with the least error, using the constraints (limits,
dependencies, locking) provided.
The time taken to fit data depends on the number of bands, the number of band
parameters, the maximum number of iterations, and the constraints (such as
limits, doublets, locking). Additionally, the time depends on how close the curve
fit bands already model the experimental data. The fit is complete when either
a) the maximum number of iterations is reached or b), convergence is reached
and the current estimates can no longer be adjusted enough to have a positive
impact on the results.
When the fit is completed, the bands and resulting synthetic curve reflect the
new estimates. If the fit has been successful, the resulting curve usually
more closely matches the experimental data. The change in data is more dramatic
if this is the first attempt at fitting using the given parameters. The
goodness-of-fit value shown in the results table reflects the error-mean-square (EMS)
value of the latest fit and provides a way to compare results from other fits
on the same or similar data.
After a fit, click the Results button for details on the new estimates and the
results of the fit.
Displaying and Interpreting Curve Fit Results
Additional Curve Fitting Topics