Re: [Help-gsl] Linear fitting

[Francisco M Neto]

Patrick,

        Thanks a lot! I really should have paid more attention to how I
handled those indices. I've adapted my code to incorporate your
suggestions and it's working now. 

Cheers,
Francisco

On Thu, 2018-05-03 at 10:47 -0600, Patrick Alken wrote:
> Francisco,
> 
>    You have some errors in the way you handle your 2D arrays. I have 
> made a small modification to define new 1D arrays (x and y) and store 
> your fit data into those arrays. The program output now says:
> 
> Performing fit with 20 points...done! (status = 0)
> ====== BET results ======
> y = 0.000022 + 0.005321 x
> [2.54531e-11, -1.26944e-10
>   -1.26944e-10, 8.07463e-10]
> Chi Square = 0.000000
> =========================
> 
> My changes are attached - you can find the lines I changed with the
> "PA" 
> keyword
> 
> On 05/03/2018 04:13 AM, Francisco M Neto wrote:
> > Greetings!
> > 
> >     I have been trying to use gsl for some linear least squares
> > fitting, but for some reason I'm getting bad results.
> > 
> >     Instead of returning the results of the fit, gsl_fit_linear()
> > returns -nan for all parameters. I'm not sure if I'm passing
> > something
> > wrong to it, but I do know that there is no problem with the data
> > (I've
> > ran it through other LS fitting procedures and results are
> > consistent).
> > 
> >     Attached are the source and an example data file. I've compiled
> > and ran with:
> > 
> > $ gcc -I/usr/include/gsl/ -lm -lgsl -lgslcblas -o teste_bet
> > teste_bet.c
> > 
> > $ ./teste_bet data.dat
> > 
> 
> 
-- 
--
[]'s,

Francisco M Neto
Institut für Anorganische Chemie
Universität Duisburg-Essen