[Scilab-users] Scilab leastsq exponential fitting

[Antoine Monmayrant]

Le 28/07/14 17:44, chloe.kykam a écrit :
> Yes, I did check the dimensions of x, ri, r and wr before, which all came
> back to be 1024x1.

OK,

Next culprit: your fit function is not defined for every values of a(1) 
and a(2).
While optimizing and calculating the numeric derivative, leastsq can try 
a(2)=0 ...
Anyway, you might solve this by 1) giving some boundaries on possible 
values for "a" and 2) calculate the Jacobian of your fit (ie partial 
derivative of your fit function with respect to a(1) and a(2).

Antoine


>
>
>
> --
> View this message in context: http://mailinglists.scilab.org/Scilab-leastsq-exponential-fitting-tp4030949p4030975.html
> Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users
>