Hello. I have a set of points that I have fit a fourth-degree polynomial to using a TF1 function (see below). But each point has an error bar associated with it so I want to, hopefully, have my function take that uncertainty into account when making the fit (so that the fit I make best represents the data given the uncertainty). How can I do that?
The uncertainty is simply a vertical error bar for each point that goes as the sqrt(y), where y is the position of the point on the y-axis, so the error is different for each point (but follows the same sqrt(y) trend).
For some background, I am pulling from 1D histograms, basically, I’m plotting the bin contents on the y-axis on a regular graph instead of a histogram. My goal is to make a fit to the data that way and then overlay the fit line to the histogram later. I realize there may be a way to do this with the histogram directly, but I decided to go this route. I left out my histogram creation, just trust that they are TH1D with 200 bins from 0.0 to 2.0, but that may or may not be useful to know.
Thanks for your help!
TGraph* g = new TGraph();
Double_t x;
Double_t y;
// create my x and y points to plot
Double_t bins = 200;
for (Int_t i = 1; i <= bins; i++) {
x = histo->GetBinCenter(i); // I pull from a 1D histogram to get my points
y = histo->GetBinContent(i);
g->SetPoint(i, x, y); // sets my points on the xy plane
}
g->Draw(“A*”);
// creates my fit function, choosing a 4th degree poly with a range from 0 to 2
TF1* f = new TF1(“f”, “pol4”, 0.0, 2.0);
g->Fit(f);
_ROOT Version: 6.16
_Platform: Ubuntu
Compiler: Not Provided