/dq_fit_library

Primary LanguageC++

DQ fit library

Python class

This python class is based on RooFit and allows to perform binned and unbinned fits. An example of fit can be found in /tutorial:

  • Generate a toy sample 
    python tutorial.py configFit.json --gen_tutorial
  • Run the fit on the toy sample 
    python tutorial.py configFit.json --fo_fit

For validation of the code use the reference:

c++ class

This class can be used to fit invariant mass distributions (signal + backgorund). To run it is necessary ROOT6. The class contains:

  • class files: DQFitter.cxx and DQFitter.h
  • function library: fit_library/
  • test macro: create_tutorial_dataset.C, tutorial.C

The test fit code performs a simple fit of a gaussian signal and a pol0 / expo background

In the root environment run with:

  • .x DQFitter.cxx+
  • .x tutorial.C

An AnalysisResults.root file will be produced containing:

  • the canvas fit the histogram and the fitting functions
  • the canvas with the ratio Data / Fit
  • the canvas with the residuals (data - background after fit)
  • the histogram with the X2 / NDF, the number of signal events and the parameters of interest (POI)

The class can use also the RooFit library. In this case the used has to create a class following the example of GausPdf.cxx. At this point the PDF can be added to the class and used for the fit.

QC macro

This macro can be used to produce control plots reading the O2 DQ tableMaker and tableReader.In addition it extracts basic J/psi basic features (mass position and width) using the DQ fit library

In the root environment run with:

  • .x DQFitter.cxx+
  • .x run_qc.C++