Fit a model for S(Q,w) to an object, with an optional background function that is also a model for S(Q,w)
If passed an array of objects, then each object is fitted independently.
For full help, read documentation for sqw object fit_sqw_sqw:
>> help sqw/fit_sqw_sqw
Differs from multifit_sqw_sqw, which fits all objects in the array simultaneously
but with independent backgrounds.
Fit several objects in succession to a given function:
>> [wout, fitdata] = fit_sqw_sqw (w, func, pin) % all parameters free
>> [wout, fitdata] = fit_sqw_sqw (w, func, pin, pfree) % selected parameters free to fit
>> [wout, fitdata] = fit_sqw_sqw (w, func, pin, pfree, pbind) % binding of various parameters in fixed ratios
With optional 'background' function added to the function
>> [wout, fitdata] = fit_sqw_sqw (..., bkdfunc, bpin)
>> [wout, fitdata] = fit_sqw_sqw (..., bkdfunc, bpin, bpfree)
>> [wout, fitdata] = fit_sqw_sqw (..., bkdfunc, bpin, bpfree, bpbind)
If unable to fit, then the program will halt and display an error message.
To return if unable to fit, call with additional arguments that return status and error message:
>> [wout, fitdata, ok, mess] = fit_sqw_sqw (...)
Additional keywords controlling which ranges to keep, remove from objects, control fitting algorithm etc.
>> [wout, fitdata] = fit_sqw_sqw (..., keyword, value, ...)
Keywords are:
'keep' range of x values to keep
'remove' range of x values to remove
'mask' logical mask array (true for those points to keep)
'select' if present, calculate output function only at the points retained for fitting
'list' indicates verbosity of output during fitting
'fit' alter convergence critera for the fit etc.
'evaluate' evaluate function at initial parameter values only, with argument check as well
'chisqr' evaluate chi-squared at the initial parameter values (ignored if 'evaluate' not set)
Example:
>> [wout, fitdata] = fit_sqw_sqw (..., 'keep', xkeep, 'list', 0)