From: Christophe Delaere <delaere_at_fynu.ucl.ac.be>

Date: Tue, 26 Apr 2005 13:45:38 +0200

Date: Tue, 26 Apr 2005 13:45:38 +0200

Hi,

If you need to fit your chi2 distribution, the Neural-Net method may be pretty usefull, especially if you don't want to give a physical meaning to the resulting function (as it seems in your case, as you just want to interpolate the points).

You may be interested by the second example there: http://www.fynu.ucl.ac.be/users/c.delaere/level2/MLP/

Cheers,

Christophe.

On Tuesday 26 April 2005 13:11, you wrote:

*> Hello Otto,*

*> My problem is highly non gaussian and the Chi2 presents 2 local minimum*

*> reflecting the ambiguity of the physics involved. I will use some hints*

*> given by Rene and see if I can manage.*

*> See you*

*> Jean-Eric*

*>*

*> --On mardi 26 avril 2005 12:43 +0200 Otto Schaile*

*>*

*> <Otto.Schaile_at_Physik.Uni-Muenchen.DE> wrote:*

*> > Hi Jean-Eric,*

*> >*

*> > I would try to fit a TF2 (2-dim function) to the x,y, chi2 values*

*> > and then plot this function. The question is if you find a simple*

*> > function. Polynomials with high enough degree fit of course*

*> > everything.*

*> > If you want I can provide you with a simple 2 dim gauss.*

*> > When plotting the function you can adjust the number of contours*

*> > and the level heights by calls to:*

*> >*

*> > f1->SetContour(Int_t nlevels, const Double_t *levels)*

*> >*

*> > Cheers*

*> > Otto*

*> >*

*> > On Tue, 26 Apr 2005, Jean-Eric Campagne wrote:*

*> >> Dear developpers,*

*> >>*

*> >> I wander if you can help me in the following task: my final goal is to*

*> >> plot Chi2 contours in a 2 dimentions x-y plot.*

*> >>*

*> >> As a starting point, I have a table of Chi2_ij values for some (x_i,*

*> >> y_j) points. The contours are the location of (x,y) points where*

*> >> Chi2(x,y) = 10.64 (for definiteness).*

*> >>*

*> >> My problem is to get the (x,y) location that fullfill the above*

*> >> constraint, for that the Chi2 should be evaluated (interpolated) between*

*> >> the different (x_i,y_j,Chi2_ij) values.*

*> >>*

*> >> How ROOT can help me in this double task: 1) perform the interpolation*

*> >> to find the (x,y) that Chi2(x,y) = 10.64 and 2) plot the result?*

*> >>*

*> >> Regards*

*> >> J.E Campagne*

*> >>*

*> >> ........................................................................*

*> >> ... .LAL - IN2P3 - CNRS*

*> >> .LAL - B.P 34 - 91898 Orsay Cedex - France*

*> >> .Piece 108*

*> >> .Tel +33 (0)1 64 46 84 29*

*> >> .Fax +33 (0)1 64 46 83 97*

*> >> ........................................................................*

*> >> ...*

*> >*

*> > --*

*> > Dr. Otto Schaile Department f"ur Physik, LMU M"unchen*

*> > Phone: +49 89 289 14070 Am Coulombwall 1, D-85748 Garching, Germany*

*> > Mobil: +49 160 1553598*

*> > FAX: +49 89 289 14072 EMail: Otto.Schaile_at_Physik.Uni-Muenchen.DE*

*>*

*> ...........................................................................*

*> .LAL - IN2P3 - CNRS*

*> .LAL - B.P 34 - 91898 Orsay Cedex - France*

*> .Piece 108*

*> .Tel +33 (0)1 64 46 84 29*

*> .Fax +33 (0)1 64 46 83 97*

*> ...........................................................................*

-- +-----------------------------------------------------------\|/---+ | Christophe DELAERE office: e253 !o o! | | UCL - FYNU phone : 32-(0)10-473234 ! i ! | | chemin du cyclotron, 2 fax : 32-(0)10-452183 `-' | | 1348 Louvain-la-Neuve BELGIUM e-mail: delaere_at_fynu.ucl.ac.be | +-----------------------------------------------------------------+Received on Tue Apr 26 2005 - 13:45:48 MEST

*
This archive was generated by hypermail 2.2.0
: Tue Jan 02 2007 - 14:45:07 MET
*