[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bayani



 Youcef Rabah Rahal wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salam,

On Wednesday 03 March16 2004 :17, Baker Abdulhaq wrote:

OK, My first C++ program. please tell me the marks.
what next.
I have only used gsl. I did nothing important.

Sorry Baker, I did not have the chance to test your code yet. I need to
install GSL first. Apart from that, the files look fine to me (I suppose you
already tested them ?).

However, in Bayani, I want to keep the dependency tree as short as possible
(for the time being at least)... So the user is asked to install only QT 3
(which is usually already there by default).

So, for the time being, we have to code a performant existing fitting
algorithm, independently from any existing C/C++ library. And when it comes
to analytical solutions (linear fit), there's nothing to do except coding the
exact least squares _expression_ ;-)

http://mathworld.wolfram.com/LeastSquaresFitting.html [straight-forward...]

We need to do that first (we have to take into account the errors on the data
too...).

What do you think ? Can you do that ? (you'll only need to change the call to
gsl_fit_linear() with a new one).
I have no problem  in coding it. But I suppose that we need not only linear regression to implement. We need to make multifit,  may be non-linear regrisions  and interpolation also. It will be good to have GSL to buld a serious mathematical program. The  advantages of using  GSL may be that,  we use a code that other people are testing, impruving  and optimizing.
However, I have fun coding. I will code it may self.

Salam,

- --
Youcef R. Rahal
Arabeyes.org
http://www.arabeyes.org/~rahal
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2. 3(GNU/Linux)

iD8DBQFAStVfHDRR6Cd0eSYRArd9AJ9rPPg6S8OsTaxE++FlhIHaTMwKpwCfd10y
k/jxVRhKiVm/t70T0Js/4YM=
=nFyD
-----END PGP SIGNATURE-----

_______________________________________________
Developer mailing list
Developer at arabeyes dot org
http://lists.arabeyes.org/mailman/listinfo/developer