Skip to content

Latest commit

 

History

History
244 lines (213 loc) · 13.3 KB

Example.md

File metadata and controls

244 lines (213 loc) · 13.3 KB
Raw

Example of a KLFitter implementation

Introduction

The purpose of the provided example is to demonstrate how to implement KLFitter in your analysis. The example code and the input files are designed to provide a minimalist input and implementation. The example also shows how to retrieve the output from KLFitter after the fitting. The ttbar lepton+jets likelihood, implemented in the class LikelihoodTopLeptonJets has been chosen for the example, but from the implementation it should be straightforward to extend the implementation to other available likelihoods. Detailed explanations of all likelihoods can be found in the main KLFitter documentation.

How to run the example

To run the example, compile the KLFitter library first. For build and installation instructions, please refer to the README file. Then run the executable, with the path to the KLFitter source directory as an additional argument. For example, if you followed the build instructions of the README, the cmake command will store the executable in the bin subdirectory. Then call

./bin/example-top-ljets.exe ../

The location of the KLFitter source directory is needed to locate the input file for the example, stored under data/examples, and the transfer functions, stored under data/transferfunctions.

Implementation

The implementation of the example consists of three different files:

  • An input ROOT file with a few hundred ttbar lepton+jets events. The file contains all necessary variables to run the KLFitter lepton+jets likelihood (described in detail in the main documentation).
  • A header file for the TreeReaderTopLJets class that provides an interface to read the variables from the input file easily. This class is used to loop over all input events and load the event information.
  • An implementation file that provides a minimalist implementation of KLFitter in a toy-analysis. The file contains detailed comments about every command and sub-step to use KLFitter.

Input file

The input file for the provided example (../data/examples/top-ljets-input.root) is a HepSim sample taken from the home page of the HepSim project [1]. The sample simulates ttbar+jet processes in proton-proton collisions at a center-of-mass energy of 13 TeV. The processes are generated using the MadGraph matrix element generator interfaced with Herwig6. The simulated detector corresponds to the Snowmass detector [2], which is a toy detector used for the 2013 Snowmass studies.

The original Delphes format of the input ROOT files is transformed into a standard ROOT tree, the events are skimmed and only variables relevant for the KLFitter reconstruction are kept. The event skimming sets the following requirements for events to be kept:

  • at least 4 jets with a transverse momentum of at least 25 GeV. The eta of the jets is required to fulfill |eta| < 4.9.
  • exactly one charged lepton (i.e. electron or muon) with a transverse momentum of at least 25 GeV and |eta| < 4.9.

Input file variables

The input ROOT file contains only one ROOT tree, nominal. The branches in this tree are described below:

  • lepton_pt: Float. Measured transverse momentum of the charged lepton (electron or muon), in GeV.
  • lepton_eta: Float. Measured eta of the charged lepton (electron or muon).
  • lepton_cl_eta: Float. Measured eta of the charged lepton (electron or muon) as measured in the calorimeters ("cluster eta") - this is needed for electrons only.
  • lepton_phi: Float. Measured phi of the charged lepton (electron or muon).
  • lepton_e: Float. Measured energy of the charged lepton (electron or muon), in GeV.
  • met_met: Float. Magnitude of the measured missing transverse momentum, in GeV.
  • met_phi: Float. Phi component of the missing transverse momentum.
  • sumet: Float. Scalar sum of the transverse energy of an event.
  • lepton_is_e: Char. A flag that carries information if the lepton in an event is an electron.
  • lepton_is_mu: Char. A flag that carries information if the lepton in an event is a muon. This is in principle redundant if lepton_is_e, but can be used for sanity checks.
  • jet_pt: A vector of floats. Each element corresponds to the measured transverse momentum of a reconstructed jet in an event, in GeV.
  • jet_eta: A vector of floats. Each element corresponds to the measured eta of a reconstructed jet in an event.
  • jet_phi: A vector of floats. Each element corresponds to the measured phi of a reconstructed jet in an event.
  • jet_e: A vector of floats. Each element corresponds to the measured energy of a reconstructed jet in an event, in GeV.
  • jet_btag_weight: A vector of floats. Each element corresponds to the b-tagging discriminant of the corresponding reconstructed jet. A dummy value (1) is used in this example.
  • jet_has_btag: A vector of chars. Each element represents a flag if the corresponding reconstructed jet is b-tagged or not.

Output of the example

The example code prints some important variables for the first processed event. The output variables are also stored in form of a ROOT file, top-ljets-output.root, which contains the following branches:

  • klf_bhad_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the b-quark assigned by KLFitter to come from the top quark that decays hadronically.
  • klf_bhad_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of the b-quark assigned by KLFitter to come from the top quark that decays hadronically.
  • klf_bhad_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the b-quark assigned by KLFitter to come from the top quark that decays hadronically.
  • klf_bhad_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the b-quark assigned by KLFitter to come from the top quark that decays hadronically.
  • klf_bhad_jet_index: A vector of unsigned integers. Each element corresponds to one permutation. Value in an element represents the position in a jet vector of the b-quark assigned by KLFitter to come from the top quark that decays hadronically.
  • klf_blep_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the b-quark assigned by KLFitter to come from the top quark that decays semi-leptonically.
  • klf_blep_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of b-quark assigned by KLFitter to come from the top quark that decays semi-leptonically.
  • klf_blep_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the b-quark assigned by KLFitter to come from the top quark that decays semi-leptonically.
  • klf_blep_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the b-quark assigned by KLFitter to come from the top quark that decays semi-leptonically.
  • klf_blep_jet_index: A vector of unsigned integers. Each element corresponds to one permutation. Value in an element represents the position in a jet vector of the b-quark assigned by KLFitter to come from the top quark that decays semi-leptonically.
  • klf_lquark1_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the first quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark1_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of the first quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark1_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the first quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark1_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the first quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark1_jet_index: A vector of unsigned integers. Each element corresponds to one permutation. Value in an element represents the position in a jet vector of the first quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark2_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the second quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark2_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of the second quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark2_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the second quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark2_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the second quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lquark2_jet_index: A vector of unsigned integers. Each element corresponds to one permutation. Value in an element represents the position in a jet vector of the second quark assigned by KLFitter to come from the W boson that decays hadronically.
  • klf_lepton_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the lepton from the leptonically decaying W boson.
  • klf_lepton_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of the lepton from the leptonically decaying W boson.
  • klf_lepton_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the lepton from the leptonically decaying W boson.
  • klf_lepton_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the lepton from the leptonically decaying W boson.
  • klf_neutrino_pt: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted pT of the neutrino from the leptonically decaying W boson.
  • klf_neutrino_eta: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted eta of the neutrino from the leptonically decaying W boson.
  • klf_neutrino_phi: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted phi of the neutrino from the leptonically decaying W boson.
  • klf_neutrino_e: A vector of floats. Each element corresponds to one permutation. Value in an element represents the fitted energy of the neutrino from the leptonically decaying W boson.
  • klf_loglikelihood: A vector of doubles. Each element corresponds to one permutation. Value in an element represents the logarithm of the KLFitter likelihood.
  • klf_event_probability: A vector of doubles. Each element corresponds to one permutation. Value in an element represents the event probability for a given permutation.
  • klf_fit_minuit_did_not_converge: A vector of chars. Each element corresponds to one permutation. Value in an element represents a flag about convergence of the fit.
  • klf_fit_aborted_to_nan: A vector of chars. Each element corresponds to one permutation. Value in an element represents a flag storing information about the fit being aborted due to a NaN value.
  • klf_fit_parameter_at_limit: A vector of chars. Each element corresponds to one permutation. Value in an element represents a flag storing information about the fit being at the boundary of at least one fit parameter.
  • klf_fit_invalid_transfer_function: A vector of chars. Each element corresponds to one permutation. Value in an element represents a flag storing information about the fit being in the region where the transfer functions are not valid.

How to identify permutations with the highest event probability

If you want to use values corresponding to permutations with, for example, the highest event probability, check the klf_event_probability vector and find the position of the element with the highest value. Store the index of this permutation in the vector. Then take the values from all variable vectors at exactly the same position. These values correspond to the permutation with the highest event probability.

References

[1] S. V. Chekanov., HepSim: a repository with predictions for high-energy physics experiments, Adv. High Energy Phys. 2015 (2015) 136093.

[2] J. Anderson et. al., Snowmass Energy Frontier Simulations, arXiv:1309.1057.