Re: TGraph2D::Fit doesn't return

Hi there,

Can you send your code with your datafile (if any) ? I would like to reproduce the problem.

John

On Fri, 2010-04-09 at 23:10 +0200, Sébastien wrote:
> Root 2.06/00b
> Linux x86
>
> Good evening,
>
> First step with ROOT, I'm trying to fit some data for my school project.
> I'm using ROOT within my c++ app.
>
> I'm using a TGraph2D, filled with points using the TGraph2D::SetPoint
> method. The graph render well when I call the Draw() function. It's a
> graph of 400x400 points.
>
> I try to fit the TGraph2D by a 2D Gaussian, like that :
>
> TF2 formula("gaus2d", Gaus2D, 0, 400, 0, 400, 5);
> formula.SetParameters(255, 200, 200, 200, 200);
> formula.SetParNames("Const","X_{0}","#sigma_{x}", "Y_{0}","#sigma_{y}");
>
> graph->Fit(formula.c_str(), "SVR");
>
> (Obviously, graph is a TGraph2D variable)
>
> The fit works well, I can see the output on the console (verbose on),
> but, when the fit converged, the app freeze. I've the following on my
> console :
>
> FCN=2040.94 FROM MIGRAD STATUS=CONVERGED 1161 CALLS 1162 TOTAL
> EDM=1.31725e-10 STRATEGY= 1 ERROR MATRIX ACCURATE
> EXT PARAMETER STEP FIRST
> NO. NAME VALUE ERROR SIZE DERIVATIVE
> 1 Const 2.54956e+02 5.05628e-02 7.90304e-04 -9.42453e-05
> 2 X_{0} 2.00000e+02 3.12950e-03 9.53674e-05 -3.34396e-03
> 3 #sigma_{x} 1.57515e+01 3.12097e-03 5.62767e-05 -5.08722e-04
> 4 Y_{0} 2.00000e+02 3.12950e-03 9.53674e-05 1.03130e-03
> 5 #sigma_{y} 1.57515e+01 3.12097e-03 5.62767e-05 2.91193e-03
> EXTERNAL ERROR MATRIX. NDIM= 25 NPAR= 5 ERR DEF=1
> 2.557e-03 -4.742e-10 -7.869e-05 -4.748e-10 -7.869e-05
> -4.742e-10 9.794e-06 2.919e-11 1.008e-13 -2.210e-14
> -7.869e-05 2.919e-11 9.740e-06 -2.060e-14 -1.764e-08
> -4.748e-10 1.008e-13 -2.060e-14 9.794e-06 2.933e-11
> -7.869e-05 -2.210e-14 -1.764e-08 2.933e-11 9.740e-06
> PARAMETER CORRELATION COEFFICIENTS
> NO. GLOBAL 1 2 3 4 5
> 1 0.70582 1.000 -0.000 -0.499 -0.000 -0.499
> 2 0.00000 -0.000 1.000 0.000 0.000 -0.000
> 3 0.57630 -0.499 0.000 1.000 -0.000 -0.002
> 4 0.00000 -0.000 0.000 -0.000 1.000 0.000
> 5 0.57630 -0.499 -0.000 -0.002 0.000 1.000
> EXTERNAL ERROR MATRIX. NDIM= 5 NPAR= 5 ERR DEF=1
> 2.557e-03 -4.742e-10 -7.869e-05 -4.748e-10 -7.869e-05
> -4.742e-10 9.794e-06 2.919e-11 1.008e-13 -2.210e-14
> -7.869e-05 2.919e-11 9.740e-06 -2.060e-14 -1.764e-08
> -4.748e-10 1.008e-13 -2.060e-14 9.794e-06 2.933e-11
> -7.869e-05 -2.210e-14 -1.764e-08 2.933e-11 9.740e-06
>
> And nothing more, I have to kill the app...
>
> It works fine if I use a 40x40 points graph, but I need a 400x400, and
> maybe even more.
>
> Is there something I've made wrong?
>
> Thanks for your answer.
>
Received on Sat Apr 10 2010 - 00:25:03 CEST