Numeric least squares fit

1 view (last 30 days)
Sergio Quesada
Sergio Quesada on 2 Oct 2018
Edited: Sergio Quesada on 2 Oct 2018
Good evening to everybody,
I need some help in this fitting...
I have a collection of experimental data (texp, Iexp). Using these data, I generate a new collection of data, (texp, rteor), by solving a system of one equation per each couple of data, using fsolve with arrayfun:
rteor= fsolve(@(r) arrayfun(@(R,T) FC.*integral(@(x) exp(-1./(30.*x.*log(x))), 1, R)-(T-t),r,texp), inpts);
(inpts=initial points; ). Then, I generate a new matrix of points, (texp, Iteor), evaluating the previous rteor data on a new function Iteor:
Iteor=rteor.*exp(a./(rteor(t).*log(rteor))) + (Io./(rteor(t).^4));
Well, now I want to fit this (Iteor, tteor) collection to (Iexp, texp) experimental collection by optimizing only the parameter "t" in the first "rteor" expression above. I would prefer to make it with a numerical method, not symbolic as the combination of
lsqcurvefit ( fsolve ( arrayfun ) )
, which I have already tried, takes a lot of time...
Hope anybody can help me. Thank you so much !!

Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!