fitting code, statistical library

hi,
i am looking for a library in C++ which enables me to fit a function to a
set of data...ideally in pure STL C++...
Please share your insight on the subject.
my thanks
TF

re: fitting code, statistical library

<OT>
Perhaps ROOT? It is not perhaps "pure STL C++" as you put it but anyway...
http://root.cern.ch

Regards,
Peter Jansson
http://jansson.net/
</OT>

re: fitting code, statistical library

well, what do you mean by "but anyway..."
is there any thing better than root (which is highly adviced by my physicist
friends)
TF
"Peter Jansson" <webmaster@jansson.net> wrote in message
news:Pine.NEB.4.60.0407172116190.23798@otaku.frees hell.org...[color=blue]
> <OT>
> Perhaps ROOT? It is not perhaps "pure STL C++" as you put it but anyway...
> http://root.cern.ch
>
> Regards,
> Peter Jansson
> http://jansson.net/
> </OT>[/color]

re: fitting code, statistical library

On Sun, 18 Jul 2004, To Forum wrote:[color=blue]
> well, what do you mean by "but anyway..."
> is there any thing better than root (which is highly adviced by my physicist
> friends)
> TF[/color]

<OT>
I meant: But anyway, if might be worth a try.

The mathematics behind "fitting functions" is logical as far as I know so
to define what is "better" and what is not so good is perhaps an issue of
calculation speed. If you know what function to fit, then perhaps the best
(the fastest with regards to calculation speed) solution is to derive your
own least squares measure of the fit and code it.

Regards,
Peter Jansson
http://jansson.net/
</OT>