Fitting in BornAgain deals with estimating the optimum parameters in the numerical model (i.e. scattering simulation)
by minimizing the difference between simulated and reference data. The features include
A variety of built in minimization algorithms.
Possibility to use external minimizers (bumps, lmfit, etc).
Choice over possible fitting parameters, their properties and correlations.
Full control over objective function calculations, including the use of different normalizations and assignments of different masks and weights to different areas of the reference data.
Possibility to fit simultaneously an arbitrary number of data sets.
The BornAgain fitting interface was redesigned in October, 2018 for better compatibility
with third party minimizers (scipy.optimize, lmfit, bumps).
In the following explanations we will be mostly focused on GISAS data. However, all of the following is also applicable for specular and off-specular fits.