For this particular example, Carlos algorithm spits out in < hart beat !
Thanks again Carlos, it will surely help revamp some/many of the GenfitMinimize.
Shina

Wrote

For this particular example, Carlos algorithm spits out in < hart beat !
Thanks again Carlos, it will surely help revamp some/many of the GenfitMinimize.
Shina

Could you be more specific please. What example are you talking about?

Regards,
Radovan

Wrote

For this particular example, Carlos algorithm spits out in

Please, share your application.
LM covers a good % of all fits in Mathcad.
Cases LM fails, Conjugate Gradient is the sole rescue.
Many model functions can only be fitted by hand.

Cheers ... Jean

For this particular example, Carlos algorithm spits out in < hart beat !
Thanks again Carlos, it will surely help revamp some/many of the GenfitMinimize.
Shina
...................
You are right About LM, however not so simple.
al_nleqsolve, LM, ConjugateGradient that produce in Smath, all dependent
upon initials and reflexivity of the model +++ the approach technique.
Data reduction is a great approach technique [Kirby, Thermocouples ...]
In Mathcad LM as well as CG, a technique that works well to initialize
wrt many models is to pass manually initials through 2 points.
That was not explored with Carlos LM.
That technique through 2 points did work well in some Smath fit,
but failed in other fits. Maybe you are working hard on that stuff ?
Attached for the interest of this visit.



Genfit Chemist.sm (34 KiB) downloaded 72 time(s).
Genfit Minimize Rational KHL.sm (45 KiB) downloaded 70 time(s).
Genfit Pharmacokinetics.sm (41 KiB) downloaded 64 time(s).

... for your delight => try Laser

Genfit Convolution Laser.sm (231 KiB) downloaded 72 time(s).