Hi all,
I am fitting a function and making its 1-sigma and 2-sigma band like following:
f1 = r.TF1("f1","pol7",600,2500)
Alpha.Fit("f1")
hConf1s = Alpha.Clone("hConf1s");
hConf2s = Alpha.Clone("hConf2s");
hConf1s.Reset();
hConf2s.Reset();
r.TVirtualFitter.GetFitter().GetConfidenceIntervals(hConf1s, 0.68);
r.TVirtualFitter.GetFitter().GetConfidenceIntervals(hConf2s, 0.95);
hConf1s.SetStats(0);
hConf2s.SetStats(0);
hConf1s.SetMarkerSize(0);
hConf2s.SetMarkerSize(0);
hConf1s.SetFillColor(r.kGreen+1);
hConf2s.SetFillColor(r.kOrange);
hConf1s.SetLineColor(r.kGreen+1);
hConf2s.SetLineColor(r.kOrange);
hConf2s.Draw("e3 same");
hConf1s.Draw("e3 same");
Alpha.Draw("E1 same");
c1.Draw()
The output plot looks like [1]. I need this function in 38 bins (for some particular reason). But, the histogram that I am fitting has poor statistics. So, to get sufficient statistics in each bin first I bin them in 15 bins only and after fitting the histogram later I am grabbing histogram in 38 bins using the fitted function.
But, how I could get the same for 1-sigma up/down? To get the information in 38 bins, first I should get a functional form only then I can get the required thing.
Please let me know how I can get the functional form of +/- 1-sigma band or if there is any other method to do this, please suggest.
Thanks & regards,
Ram
[1]
ROOT Version (e.g. 6.12/02): 6.12/06
Platform, compiler (e.g. CentOS 7.3, gcc6.2): linuxx8664gcc