[Scilab-users] Fwd: How to get fjac from lsqrsolve

[Stéphane Mottelet]

Hello,

You can use the numderivative macro, with your residual function as 
first argument.

hth

S.


Le 17/09/2015 16:18, Pablo Caron a écrit :
> Dear user group,
>
> I need to extract the Jacobian from the lsqrsolve function. The 
> fortran routine lmdif (http://www.netlib.org/minpack/lmdif.f) provide 
> this as the output fjac, but lsqrsolve does not have this option.
>
> I need to evaluate the goodness of a regression I made using lsqrsolve 
> and I found that this is done through the correlation matrix. I need 
> either the correlation matrix or the jacobian.
>
> I would like to get an output similar to the following
>
> Final set of parameters Asymptotic Standard Error
> ======================= ==========================
> m = 0.174844      +/- 0.0002964 (0.1695%)
> b = 1506.34     +/- 5.756 (0.3821%)
>
> correlation matrix of the fit parameters:
>        m     b
> m  1.000
> b -0.448 1.000
>
> Regards
>
> Pablo
>
>
>
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users


-- 
Département de Génie Informatique
EA 4297 Transformations Intégrées de la Matière Renouvelable
Université de Technologie de Compiègne -  CS 60319
60203 Compiègne cedex

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20150917/3f1595fb/attachment.htm>