Hello,
I want to calculate the integral error of an fitted function in a bin range. As I have read in a tutorial it should not be necessary to use a covariance matrix. But I get only the message: “Last used Fitter has different parameter values”. Here is my code example:
TF1* fitfunc = new TF1("boltzmann","[0] * exp(-x/[1])"); //expected fit function
histo_Projection[i] = histo->ProjectionY(Form("histo_Projection_%i",i),i,i); //embedded in an loop over "i" to make 1D histograms from 2D histogram called "histo"
histo_Projection[i] -> Fit(fitfunc,"","",0.1,1.0);
TH1F* h_fitfunc = (TH1F*) fitfunc->GetHistogram(); //create an histogram from function
IntegralError[i] = 0;
for(int j=0; j<=h_fitfunc->GetNbinsX(); j++){ //loop over entries
IntegralError[i] = IntegralError[i] + fitfunc->IntegralError(h_fitfunc->GetBinCenter(j),h_fitfunc->GetBinCenter(j+1) ); //sum up the error in bin ranges necessary
}
Can someone help me to fix the problem?
Thanks a lot,
Johannes