Electron Identification Validation

Complete: 3

Contents:

Contacts

  • Pascal Vanlaer
  • David Futyan
  • Roberto Salerno
  • Pedro Ribeiro

Goals of this page

To summarize the information concerning the validation in CMSSW of the cut based method of Electron Identification, developed during PTDR analysis, for benchmark physics channels.

Validation code

Electron ID validation code available in CMSSW CVS area EgammaAnalysis/Examples

Required Packages

   EgammaAnalysis/ElectronIDProducers
   EgammaAnalysis/ElectronIDAlgos
   AnalysisDataFormats/Egamma
 

How to install

   cvs co EgammaAnalysis/Examples
   scramv1 b

How to run

  cd EgammaAnalysis/Examples/
  cmsRun test/testElectronIDValAnalyzer.cfg 
for efficiency/purity of electron reconstruction and identification

or

  cd EgammaAnalysis/Examples/
  cmsRun test/testElectronIDFakeRateAnalyzer.cfg 
for electron fake rate

The test configuration files testElectronIDValAnalyzer.cfg and testElectronIDFakeRateAnalyzer.cfg include in the sequence of the CMSSW process the electron identification step. The identification cuts are completely defined in EgammaAnalysis/ElectronIDProducers/data/ptdrElectronId.cfi. As input for the analysis step the user needs to specify the tags of the EDM collections for reconstructed electrons, GEN particles and the association map between the reconstructed electrons and their identification. The default labels are stated in data/ElectronIDValAnalyzer.cfi.

The output is a ROOT file containing several histograms. The average efficiency and purity of the electron reconstruction and identification are stored also in the ROOT file as branches of the TTree named "EfficiencyInf". The name of the ROOT file, the number of bins and the range of the histograms are specified in the cfi files data/ElectronIDValAnalyzer.cfi and data/ElectronIDFakeRateAnalyzer.cfi . In EgammaAnalysis/Examples/test/ there are two root macros ElectronIDValAnalyzer.C and ElectronIDFakeRateAnalyzer.C that can be used to format the histrograms stored in the ROOT files.

Results of validation of electron identification

In this section results on efficiency of electron identification are shown for five physics datasets : Z->ee, W->e nu_e, Di-electrons (pt 5-100 GeV), Z'->ee and g1g1->4e (1/R=500 GeV). The cut based method was applied, using the loose,medium and tight criteria. The cut values are defined in CVS . For reference we list the cut values for golden and showering electrons in the barrel and in the endcap. Note that the identifications cuts were developed independently of each other, i.e., tight is not necessarily a constrained subset of medium.

Golden electrons in the barrel

Cut loose medium tight
E/p in [0,inf] [0.9,1.3] [0,inf]
E/p out [0.6,2.5] [0.6,2.5] [0.6,inf]
!CMS.DeltaEtaIn [0,0.008] [0,0.004] [0,0.004]
!DeltaPhiIn [0,0.06] [0,0.04] [0,0.02]
!DeltaPhiOut [0,0.011] [0,0.011] [ 0,0.02]
H/E [0,0.09] [0,0.06] [0,0.05]
E9/E25 [0.7,1] [0.7,1] [0.8,1]
sigmaEtaEta [0,inf] [0.005,0.011] [0.005,0.011]
sigmaPhiPhi [0,inf] [0.005,0.015] [0,inf]

Showering electrons in the barrel

Cut loose medium tight
E/p in [0,inf] [0.6,inf] [0,inf]
E/p out [0.5,inf] [0.75,inf] [0.75,inf]
!CMS.DeltaEtaIn [0,0.009] [0,0.007] [0,0.005]
!DeltaPhiIn [0,0.08] [0,0.08] [0,0.04]
!DeltaPhiOut [0,inf] [0,inf] [ 0,inf]
H/E [0,0.12] [0,0.14] [0,0.05]
E9/E25 [0.5,1] [0,1] [0.65,1]
sigmaEtaEta [0,inf] [0.005,0.011] [0.005,0.011]
sigmaPhiPhi [0,inf] [0,inf] [0,inf]

Golden electrons in the endcap

Cut loose medium tight
E/p in [0,inf] [0.9,inf] [0,inf]
E/p out [0.6,2.5] [0.6,2.0] [0.5,inf]
!CMS.DeltaEtaIn [0,0.008] [0,0.007] [0,0.005]
!DeltaPhiIn [0,0.06] [0,0.06] [0,0.04]
!DeltaPhiOut [0,0.02] [0,0.02] [ 0,0.02]
H/E [0,0.09] [0,0.1] [0,0.07]
E9/E25 [0.8,1] [0.85,1] [0.8,1]
sigmaEtaEta [0,inf] [0.008,0.022] [0.008,0.030]
sigmaPhiPhi [0,inf] [0.00,0.02] [0,inf]

Showering electrons in the endcap

Cut loose medium tight
E/p in [0,inf] [0.7,inf] [0,inf]
E/p out [0.5,inf] [0.8,inf] [0.8,inf]
!CMS.DeltaEtaIn [0,0.009] [0,0.008] [0,0.005]
!DeltaPhiIn [0,0.09] [0,0.07] [0,0.05]
!DeltaPhiOut [0,inf] [0,inf] [ 0,inf]
H/E [0,0.12] [0,0.12] [0,0.07]
E9/E25 [0.5,1] [0,1] [0.65,1]
sigmaEtaEta [0,inf] [0.0,0.3] [0.008,0.022]
sigmaPhiPhi [0,inf] [0,inf] [0,inf]

  • Reco electrons taken from PixelMatchGsfElectron candidates
    • electron preselection

  • GEN-RECO matching criteria for efficiency and purity calculations
    • a reconstructed electron matches a Gen electron if and has same charge
    • if more than one reconstructed electrons satisfies these conditions, the one with closest relative to Gen electron is selected
    • only prompt electrons are considered

  • for each dataset, the reconstruction and identification efficiencies are plotted as a function of the , and of the Gen electron. The reconstruction efficiency represents the probability of a Gen electron which passes the preselection to be reconstructed as a PixelMatchGsfElectron candidate. The identification efficiency represents the probability of the a Gen electron which was reconstructed to pass the electron identification cuts.

  • for each dataset, tables are shown with the average efficiency and purity of electron reconstruction and identification, defined as:
    • Reconstruction efficiency = # matched reco electrons / # Gen electrons
    • Efficiency of identification = # matched good electrons / # reco matched electrons
    • Efficiency of reconstruction and identification = # matched good electrons / # Gen electrons
    • Purity of reconstruction = # matched reco electrons / # reco electrons
    • Purity of identification = # matched good electrons / # good electrons

Z->ee (CMSSW_1_6_0) Final State : 2 electrons

  • dataset /RelValZEE/CMSSW_1_6_0-RelVal-1188837971/GEN-SIM-DIGI-RECO 5200 evts

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Electron Purity as a function of Reco Electron and

W->e nu_e (CMSSW_1_6_0) Final State : 1 electron + missing energy

  • dataset /RelValWE/CMSSW_1_6_0-RelVal-1188837971/GEN-SIM-DIGI-RECO 5200 evts

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Electron Purity as a function of Reco Electron and

Di electron (CMSSW_1_6_0) Final state : 2 electrons with flat pt spectrum between 5 and 100 GeV, no proton-proton Underlying Event (sample generated with Particle Gun)

  • dataset /RelValDiElectron_Pt_5_100/CMSSW_1_6_0-RelVal-1190216041/GEN-SIM-DIGI-RECO 25500 evts

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Electron Purity as a function of Reco Electron and

Z'->ee (CMSSW_1_6_0) Z Prime mass = 1 TeV ; Final state : 2 very energetic electrons with ~ 360 GeV

  • dataset /RelValZPrimeEEM1000/CMSSW_1_6_0-RelVal-1189471912/GEN-SIM-DIGI-RECO 5025 evts

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Electron Purity as a function of Reco Electron and

UED g1g1->4e ; 1/R = 500 GeV (CMSSW_1_5_2) Final state : 4 electron + 4 gluon jets + missing energy

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Reconstruction and Identification efficiency as a function of Gen Electron

Electron Purity as a function of Reco Electron and

The following plot shows the fraction of UED events with a least i electrons from the hard process reconstructed and identified, with _i_=1,2,3,4.

InclusiveN.gif

Results for average electron efficiency and purity per electron

Efficiency

Efficiency Reconstruction Id loose Id medium Id tight Rec&Id loose Rec&Id medium Rec&Id tight
Z->ee 0.85 0.98 0.88 0.85 0.83 0.75 0.72
W->e vu_e 0.84 0.99 0.88 0.84 0.83 0.74 0.71
Di e 0.85 0.99 0.89 0.87 0.84 0.75 0.74
Z'->ee 0.82 0.99 0.91 0.96 0.81 0.75 0.79
UED 0.81 0.96 0.84 0.74 0.78 0.68 0.60

Purity

Purity Reconstruction Id loose Id medium Id tight
Z->ee 0.89 0.92 0.94 0.96
W->e vu_e 0.83 0.88 0.93 0.95
Di e 0.96 0.96 0.96 0.96
Z'->ee 0.85 0.90 0.93 0.94
UED 0.79 0.88 0.94 0.95

Fraction of charge mismatched electrons

In the following plots it is shown the distribution of charged mismatched reco electrons, defined as # reco electrons which match gen electrons geometrically and correct charge divided by # reco electrons which match gen electrons geometrically, as function of gen electron pt, eta and phi. The dataset Z->ee, processed with CMSSW_1_6_0 was analyzed.

  • dataset /RelValZEE/CMSSW_1_6_0-RelVal-1188837971/GEN-SIM-DIGI-RECO 5200 evts

Zee_RecandIDMisCharge_Pt.gif
Zee_RecandIDMisCharge_Eta.gif
Zee_RecandIDMisCharge_Phi.gif

Reconstruction and Identification efficiency for well measured electrons

In the following plots it is shown the distribution for the efficiency of reconstruction and identification of well measured electrons, as function of gen electron pt, eta and phi. Well measured electrons are reconstructed electrons with eta and phi compatible with gen electrons, correct charge and (E_reco-E_gen)/E_gen < 8% (18%) for the ECAL barrel(endcap). The dataset Z->ee, processed with CMSSW_1_6_0 was analyzed.

  • dataset /RelValZEE/CMSSW_1_6_0-RelVal-1188837971/GEN-SIM-DIGI-RECO 5200 evts

Electron fake rate from QCD jets

To estimate the electron fake rate from QCD jets, the following datasets were analyzed:

  • /QCD_Pt_20_30/CMSSW_1_5_2-CSA07-2049/GEN-SIM-DIGI-RECO
  • /QCD_Pt_30_50/CMSSW_1_5_2-CSA07-2049/GEN-SIM-DIGI-RECO
  • /QCD_Pt_50_80/CMSSW_1_5_2-CSA07-2049/GEN-SIM-DIGI-RECO
  • /QCD_Pt_80_120/CMSSW_1_5_2-CSA07-2049/GEN-SIM-DIGI-RECO

Jets reconstructed with Iterative Cone (size = 0.5) algorithm and eT > 10 GeV were considered. Jet corrections were not considered.

No electron isolation was applied

Reconstruction and Identification

Here the fake rate is defined as # reco jets that match a reconstructed and identified electron divided by # reco jets.

fakerate_div_pt.gif
fakerate_div_eta.gif

Identification only

Here the fake rate is defined as # identified electrons that match a reco jet divided by # reco electrons that match a reco jet.

fakerate_div_pt_rel.gif
fakerate_div_eta_rel.gif

Sources of information

Review Status

Editor/Reviewer and date Comments
PedroRibeiro - 06 Nov 2007 page author
JennyWilliams - 19 Dec 2007 moved page to SWGuide

Responsible: PedroRibeiro
Last reviewed by: never reviewed

Topic attachments
I Attachment History Action Size Date Who Comment
GIFgif Die_IdEta.gif r1 manage 7.5 K 2007-10-16 - 17:05 PedroRibeiro die
GIFgif Die_IdPhi.gif r1 manage 7.9 K 2007-10-16 - 17:07 PedroRibeiro Die
GIFgif Die_IdPt.gif r1 manage 6.6 K 2007-10-16 - 17:05 PedroRibeiro Die
GIFgif Die_Inclusive.gif r1 manage 5.9 K 2007-10-16 - 17:09 PedroRibeiro Die
GIFgif Die_PurEta.gif r1 manage 6.6 K 2007-10-16 - 17:08 PedroRibeiro Die
GIFgif Die_PurPhi.gif r1 manage 6.9 K 2007-10-16 - 17:09 PedroRibeiro Die
GIFgif Die_PurPt.gif r1 manage 6.3 K 2007-10-16 - 17:07 PedroRibeiro Die
GIFgif Die_RecAndIdEta.gif r1 manage 9.5 K 2007-10-16 - 17:02 PedroRibeiro DiE
GIFgif Die_RecAndIdPhi.gif r1 manage 10.4 K 2007-10-16 - 17:03 PedroRibeiro DiE
GIFgif Die_RecAndIdPt.gif r1 manage 7.8 K 2007-10-16 - 17:00 PedroRibeiro DiE
GIFgif IdEta.gif r1 manage 9.9 K 2007-10-10 - 18:57 PedroRibeiro ued
GIFgif IdPhi.gif r1 manage 10.3 K 2007-10-10 - 18:59 PedroRibeiro ued
GIFgif IdPt.gif r1 manage 6.4 K 2007-10-10 - 18:55 PedroRibeiro ued
GIFgif InclusiveN.gif r1 manage 6.2 K 2007-10-10 - 18:59 PedroRibeiro ued
GIFgif PurityEta.gif r1 manage 10.6 K 2007-10-10 - 19:00 PedroRibeiro ued
GIFgif PurityPT.gif r1 manage 6.7 K 2007-10-10 - 19:00 PedroRibeiro ued
GIFgif PurityPhi.gif r1 manage 10.7 K 2007-10-10 - 19:01 PedroRibeiro ued
GIFgif WE_IdEta.gif r1 manage 9.9 K 2007-10-16 - 15:56 PedroRibeiro WE
GIFgif WE_IdPhi.gif r1 manage 10.9 K 2007-10-16 - 15:57 PedroRibeiro WE
GIFgif WE_IdPt.gif r1 manage 6.5 K 2007-10-16 - 13:53 PedroRibeiro WE
GIFgif WE_Inclusive.gif r1 manage 5.4 K 2007-10-16 - 16:59 PedroRibeiro WE
GIFgif WE_PurEta.gif r1 manage 10.8 K 2007-10-16 - 16:58 PedroRibeiro WE
GIFgif WE_PurPhi.gif r1 manage 12.0 K 2007-10-16 - 16:58 PedroRibeiro WE
GIFgif WE_PurPt.gif r1 manage 6.7 K 2007-10-16 - 16:57 PedroRibeiro WE
GIFgif WE_RecandIdEta.gif r1 manage 12.4 K 2007-10-16 - 13:51 PedroRibeiro WE
GIFgif WE_RecandIdPhi.gif r1 manage 13.6 K 2007-10-16 - 13:52 PedroRibeiro WE
GIFgif WE_RecandIdPt.gif r1 manage 7.8 K 2007-10-16 - 13:48 PedroRibeiro WE
GIFgif WeEta.gif r1 manage 8.3 K 2007-09-24 - 17:37 PedroRibeiro  
GIFgif WePT.gif r1 manage 6.7 K 2007-09-24 - 17:37 PedroRibeiro  
GIFgif WePhi.gif r1 manage 9.0 K 2007-09-24 - 17:38 PedroRibeiro  
GIFgif Zee_IDEne_Eta.gif r1 manage 9.2 K 2007-11-26 - 15:11 PedroRibeiro Zee
GIFgif Zee_IDEne_PT.gif r1 manage 6.5 K 2007-11-26 - 15:10 PedroRibeiro Zee
GIFgif Zee_IDEne_Phi.gif r1 manage 9.8 K 2007-11-26 - 15:12 PedroRibeiro Zee
GIFgif Zee_IdEta.gif r1 manage 9.2 K 2007-10-16 - 13:39 PedroRibeiro Zee
GIFgif Zee_IdPhi.gif r1 manage 9.9 K 2007-10-16 - 13:42 PedroRibeiro Zee
GIFgif Zee_IdPt.gif r1 manage 6.5 K 2007-10-16 - 13:41 PedroRibeiro Zee
GIFgif Zee_Inclusive.gif r1 manage 5.7 K 2007-10-16 - 13:43 PedroRibeiro Zee
GIFgif Zee_PurEta.gif r1 manage 9.3 K 2007-10-16 - 13:46 PedroRibeiro Zee
GIFgif Zee_PurPhi.gif r1 manage 10.0 K 2007-10-16 - 13:47 PedroRibeiro Zee
GIFgif Zee_PurPt.gif r1 manage 6.5 K 2007-10-16 - 13:46 PedroRibeiro Zee
GIFgif Zee_RecAndIdEta.gif r1 manage 11.9 K 2007-10-16 - 13:42 PedroRibeiro Zee
GIFgif Zee_RecAndIdPT.gif r1 manage 7.6 K 2007-10-16 - 13:44 PedroRibeiro Zee
GIFgif Zee_RecAndIdPhi.gif r1 manage 13.0 K 2007-10-16 - 13:45 PedroRibeiro Zee
GIFgif Zee_RecandIDEne_Eta.gif r1 manage 11.5 K 2007-11-05 - 19:48 PedroRibeiro Zee
GIFgif Zee_RecandIDEne_PT.gif r1 manage 7.1 K 2007-11-05 - 19:48 PedroRibeiro Zee
GIFgif Zee_RecandIDEne_Phi.gif r1 manage 12.6 K 2007-11-05 - 19:48 PedroRibeiro Zee
GIFgif Zee_RecandIDMisCharge_Eta.gif r1 manage 13.2 K 2007-11-05 - 19:51 PedroRibeiro Zee
GIFgif Zee_RecandIDMisCharge_Phi.gif r1 manage 15.4 K 2007-11-05 - 19:51 PedroRibeiro Zee
GIFgif Zee_RecandIDMisCharge_Pt.gif r1 manage 7.8 K 2007-11-05 - 19:50 PedroRibeiro Zee
GIFgif Zprime_IdEta.gif r1 manage 8.6 K 2007-10-16 - 17:22 PedroRibeiro Zprime
GIFgif Zprime_IdPhi.gif r1 manage 8.9 K 2007-10-16 - 17:23 PedroRibeiro Zprime
GIFgif Zprime_IdPt.gif r1 manage 10.8 K 2007-10-16 - 17:14 PedroRibeiro Zprime
GIFgif Zprime_Inclusive.gif r1 manage 5.9 K 2007-10-16 - 17:50 PedroRibeiro Zprime
GIFgif Zprime_PurEta.gif r1 manage 9.8 K 2007-10-16 - 17:48 PedroRibeiro Zprime
GIFgif Zprime_PurPhi.gif r1 manage 10.2 K 2007-10-16 - 17:49 PedroRibeiro Zprime
GIFgif Zprime_PurPt.gif r1 manage 10.1 K 2007-10-16 - 17:48 PedroRibeiro Zprime
GIFgif Zprime_RecAndIdEta.gif r1 manage 11.0 K 2007-10-16 - 17:12 PedroRibeiro Zprime
GIFgif Zprime_RecAndIdPT.gif r1 manage 14.4 K 2007-10-16 - 17:10 PedroRibeiro Zprime
GIFgif Zprime_RecAndIdPhi.gif r1 manage 11.6 K 2007-10-16 - 17:13 PedroRibeiro Zprime
GIFgif fakerate_div_eta.gif r1 manage 11.4 K 2007-11-05 - 19:54 PedroRibeiro QCD
GIFgif fakerate_div_eta_rel.gif r1 manage 11.7 K 2007-11-05 - 19:54 PedroRibeiro QCD
GIFgif fakerate_div_pt.gif r1 manage 8.2 K 2007-11-05 - 19:53 PedroRibeiro QCD
GIFgif fakerate_div_pt_rel.gif r1 manage 7.7 K 2007-11-05 - 19:53 PedroRibeiro QCD
GIFgif recAndIdEta.gif r1 manage 12.3 K 2007-10-10 - 18:47 PedroRibeiro ued
GIFgif recAndIdPT.gif r1 manage 7.5 K 2007-10-10 - 18:55 PedroRibeiro ued
GIFgif recAndIdPhi.gif r1 manage 13.1 K 2007-10-10 - 18:47 PedroRibeiro ued
GIFgif ued500Eta.gif r1 manage 7.8 K 2007-09-24 - 19:16 PedroRibeiro  
GIFgif ued500Nelectrons.gif r1 manage 5.4 K 2007-09-24 - 19:18 PedroRibeiro  
GIFgif ued500PT.gif r1 manage 6.1 K 2007-09-24 - 19:16 PedroRibeiro  
GIFgif ued500Phi.gif r1 manage 8.1 K 2007-09-24 - 19:17 PedroRibeiro  
GIFgif ued500results.gif r1 manage 21.8 K 2007-09-24 - 14:07 PedroRibeiro  
GIFgif weresults.gif r1 manage 23.5 K 2007-09-24 - 14:08 PedroRibeiro  
GIFgif zeeEta.gif r1 manage 7.8 K 2007-09-24 - 17:36 PedroRibeiro  
GIFgif zeePT.gif r1 manage 6.4 K 2007-09-24 - 17:36 PedroRibeiro  
GIFgif zeePhi.gif r1 manage 8.5 K 2007-09-24 - 17:36 PedroRibeiro  
GIFgif zeeresults.gif r1 manage 22.4 K 2007-09-24 - 14:07 PedroRibeiro  
Edit | Attach | Watch | Print version | History: r26 < r25 < r24 < r23 < r22 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r26 - 2007-12-21 - JennyWilliams



 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback