Hi Nicolas,
Thanks for your suggestion to use the error bars of the first/last point
when computing the fit range. It was easy to implement. Now in CVS.
Rene Brun
On Fri, 13 Sep 2002, Nicolas de Sereville wrote:
> Hi Rooters
>
> When fitting TGraphErrors with a predefined function (e.g. "pol1") the
> range fitted seems to be by default [first point; last point of
> TGraphErrors].In the case of x error bars, it doesn't seem to take into
> account the left part of the error bar for the first point and the right
> part of the error bar for the last point.
> If the range of the "pol1" function is extended manually to include all
> the x error bars, then the fit is improved.
>
> Is it an intented behaviour? if not, would it be possible to change the
> default range in order to include all the x error bars with something like :
> low range = first point - 2*xerrorbar
> up range = last point + 2*xerrorbar
>
> Here is a small script which intends to reproduce this feature.
>
> Thanks in advance
>
> Nicolas
>
> P.S : root 3.02/03 on RH7.1
>
>
> void test()
> {
> gStyle->SetOptFit(1);
>
> Double_t x[3] = {1, 2, 3};
> Double_t ex[3] = {0.1, 0.1, 0.1};
> Double_t y[3] = {1, 2.1, 2.9};
> Double_t ey[3] = {0.1, 0.1, 0.1};
>
> TGraphErrors *gre = new TGraphErrors(3, x, y, ex, ey);
> gre->Draw("A*");
>
> gre->Fit("pol1");
>
> TF1 *f1 = new TF1("f1", "pol1", 0, 4);
> // gre->Fit("f1", "RV+");
> }
>
>
> --
> :-------------------------------:
> : Nicolas DE SEREVILLE :
> : C.S.N.S.M. :
> : Batiment 104 :
> : 91405 ORSAY Campus :
> : Tel : (+ 33) (0)1 69 15 48 57 :
> : Fax : (+ 33) (0)1 69 15 50 08 :
> :-------------------------------:
>
>
This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:51:09 MET