ytatus94 / nuhm2

This is the framework for NUHM2 SR optimization.

NUHM2

This is the framework for NUHM2 SR optimization.

v27. Add mee and pT requirements

1. yt_optimization.h:
   • Add three variables pTl1, pTl2, and mee
   • Add three set functions for new variables.
2. yt_optimization.cxx: Add mee and pT requirements.
3. ytEventSelection.cxx: pass pTl1, pTl2, and mee to m_optimization

v26. Change the bin of the histograms

1. yt_optimization.h: Remove vec_signal_jets_with_pt_cut and vec_signal_bjet_with_pt_cut.
2. yt_optimization.cxx:
   • Comment the cout
   • Change the constant values
   • Define new constant met_over_meff_cut
   • Remove unused code.
3. ytOptimization_plots.C:
   • Change the constant values
   • Change the bin of the histograms
   • Add the grid on the plots.

v25. Fix bin problem in plot making script

1. yt_optimization.h: Change the jet_pt_cuts value to 1
2. yt_optimization.cxx: Modify method 1, it is similar to apply_signal_region_cuts() now.
3. ytOptimization_plots.C: Correct bin to bin + 1

v24. Modify the optimization scripts

1. ytOptimization_plots.C:
   • Change the constant arraies.
   • Change the input path
   • Change the value of background_files array
   • Use m1/2=500 as input signal_file
   • Remove unused derivation_stat_weights and cross_section_kfactor_efficiency histograms
   • Add unweighted signal yields and background yields
   • Add new histogram names.
   • Modify the n_xbin and n_ybin to correct values.
   • Add 4 new histograms to keep signal and background yields and unyields.
   • Use the correct x and y labels for the historams.
   • Correct the type of the weighted yields for signal and background.
   • weighted signal yields need to multiply the signal_cross_section_kfactor_efficiency
   • Use the weighted yields to calculate the significance.
   • set significance to 0.02 and 0.01 when n_signal_weighted < 2 and n_background_weighted < 1, respectively.
2. ytOptimization_plots_method2.C
   • Change the constant arraies and constant values.
   • Change the input path
   • Change the value of background_files array
   • Use m1/2=500 as input signal_file
   • Modify the n_xbin and n_ybin to correct values.
   • Add 4 new histograms to keep signal and background yields and unyields.
   • Use the correct x and y labels for the historams.
   • Correct the type of the weighted yields for signal and background.
   • weighted signal yields need to multiply the signal_cross_section_kfactor_efficiency
   • Use the weighted yields to calculate the significance.
   • set significance to 0.02 and 0.01 when n_signal_weighted < 2 and n_background_weighted < 1, respectively.

v23. Modify method 1 and 2 to compare 4topSM results.

1. yt_optimization.h:
   • Comment vec_signal_jets_with_pt_cut and vec_signal_bjet_with_pt_cut.
   • Remove fill_Nbjets_pT()
   • Change the return type of fill_vec_jets_with_pT_cut() to int.
   • Change the number of elements for N_lept_cuts, N_bjet_cuts, N_jets_cuts, and bjet_pt_cuts.
2. yt_optimization.cxx:
   • Change the constant arraies
   • Modify method 1
   • Modify method 2
   • Modify the parts use vec_signal_jets_with_pt_cutandvec_signal_bjet_with_pt_cut`.

v22. Correct the method 1

1. yt_optimization.cxx
   • Change the meff_cuts values
   • Change the constants' value
   • Comment the output tree
   • Comment the debug_print()
   • Modifyt the method 1
   • Comment the cout in method 2
   • Add the SR regions for applying apply_signal_region_cuts()
2. ytOptimization_plots_method2.C
   • Change the input path
   • Add histograms to keep number of signal and number of backround w/ and w/o weighted.

v21. Modify the method2 plot making script.

1. ytOptimization_plots_method2.C
   • Change the input path
   • Uncomment the NUHM2 signal_cross_section_kfactor_efficiency
   • Change the declaration of output histogram
   • Set the significance to 0.02 and 0.01 when n_signal < 2 and n_background < 1, respectively.

v20. This version can get almost the same 4topSM SR events

1. yt_optimization.h:
   • Add weight data member
   • Add two new vectors vec_signal_jets_with_pt_cut and vec_signal_bjet_with_pt_cut to keep jets and b-jets with pT requirement.
   • Add two histograms h_events_survived and h_events_survived_weighted to keep the events after SR cut.
   • Add two data members n_bjet_pTX and n_jets_pTX to keep number of jets and b-jets events with pT requirement.
   • Change the type of events_survived_weighted from int to double
   • Add new method fill_vec_jets_with_pT_cut()
   • Add new argument in apply_signal_region_cuts()
   • The method debug_print() now needs to pass argument
   • Remove vec_N_bjet_pT_greater_than_some_value, get_events_survived(), and get_events_survived_weighted()
2. yt_optimization.cxx:
   • Initialize n_bjet_pTX and n_jets_pTX to zero and add branches for them.
   • Initialize two histograms h_events_survived and h_events_survived_weighted
   • Delete two branches for events_survived and events_survived_weighted
   • Add two for loops to calculate n_bjet_pTX and n_jets_pTX.
   • Modify method2 using n_bjet_pTX and n_jets_pTX
   • Midify apply_signal_region_cuts() and debug_print()
3. SR1b1 events=253, events_weighted=0.386407 SR1b2 events=408, events_weighted=0.617939 SR3b1 events=342, events_weighted=0.516968 SR3b2 events=327, events_weighted=0.510207
   • The events=342 in SR3b1 is different from Othmane's result (340)
   • All the weighted events are different from Othmane's results but very close.
   • It might because Othmane used different luminosity (36.47)

v19. Modify method2

1. yt_optimization.cxx:
   • Modify the code when N_bjet > 0
   • jet pT > jet pT cut

v18. Add event number and run number information

1. ytEventSelection.cxx: pass EventNumber and RunNb (or random_run_number) to optimization
2. yt_optimization.h:
   • Add run_number, event_number variable and vec_N_bjet_pT_greater_than_some_value.
   • Add new methods set_run_number(), set_event_number(), and fill_Nbjets_pT()
3. yt_optimization.cxx:
   • Initialize run_number, event_number, and luminosity to zero.
   • Cout run_number and event_number
   • implement fill_Nbjets_pT()
   • B-jet pT > b-jet pT cut (not >=)

v17. Modify method 2

1. yt_optimization.cxx::
   • Define constant N_lepts, N_bjets, Bjet_pT, and N_jets
   • Use the above constant in method 2
2. ytOptimization_plots.C: Plot cosmatic
3. Use the sam MET and Meff as what Chris did in yt_optimization.cxx, ytOptimization_plots.C, and ytOptimization_plots_method2.C
4. yt_optimization.h: Change the number of elements in MET and Meff arrays definition

v16. Set unreasonable significance to 0

1. yt_optimization.cxx:
   • Comment out the branches to reduce root file size
   • Change n_lept_cuts to N_lept_cuts
   • Change n_bjet_cuts to N_bjet_cuts
   • Change n_jets_cuts to N_jets_cuts
   • Uncomment method 1 and comment method 2
2. ytOptimization_plots.C:
   • Use signal_cross_section_kfactor_efficiency value from SUSYTools.
   • Modify hist_title.
   • Comment unused signal_derivation_stat_weights, signal_cross_section_kfactor_efficiency, background_derivation_stat_weights, and background_cross_section_kfactor_efficiency.
   • Add some new couts
   • Set significance to 0 if it is less then zero or reater then 100.
   • Plot cosmatic

v15. Add new script for method2

1. ytOptimization_plots_method2.C: New script
2. ytOptimization_plots.C: change the type of luminosity to float
3. yt_optimization.cxx:
   • Remove unused branches
   • Calculate lepton1_pT, lepton2_pT, jet1_pT, and jet2_pT
   • Add some couts in method2 for loop

v14. Add new method 2

1. ytEventSelection.cxx: derivation_stat_weights becomes an input of optimization
2. yt_optimization.h:
   • Add TH1F histogram h_met_over_meff
   • Add TH3F histograms h_method2_yields and h_method2_yields_weighted
   • Add new TTree leafs
   • Change the cuts type form const int to const float because TH3F uses float
   • Add binnings vector.
   • Don't fill histogram in set_derivation_stat_weights()
   • Add new method set_binning_default()
3. yt_optimization.cxx:
   • Change the cuts type form const int to const float.
   • Initialize new leafs to 0 or empty the vector.
   • Add new branches for new leafs.
   • Fill h_derivation_stat_weights after it is created.
   • Initialize new histograms h_method2_yields, h_method2_yields_weighted
   • Correct the weight calculation.
   • Calculate new variable met_over_meff
   • Comment out method 1 and add method 2.
   • Implement set_binning_default()

v13. Change the weight calculation

1. ytEventSelection.cxx: pass luminosity to optimization
2. yt_optimization.cxx: Change the calculation of weight
3. ytOptimization_plots.C: Comment the weight calculation because the root file contains correct weight now.

v12. Fix typo

1. run_bkg_samples.sh: Use lllvjj_EW6 (not llvvjj_EW6)
2. cutflow.cxx: Use lllvjj_EW6 (not llvvjj_EW6)

v11. New script to merge the input and modify script to run signal and background.

1. merge_bkg_samples.py: New script
2. run_bkg_samples.sh: Re-direct output
3. run_signal_samples.sh: Re-direct output
4. cutflow.cxx: Change the input path

v10. Use mu20_mu8noL1 trigger

1. ytEventSelection.cxx: Use mu20_mu8noL1 for v44 (not mu22_mu8noL1)
2. yt_cutflows.cxx: Use mu20_mu8noL1
3. The cutflow results agree

v09. Cutflow agree

1. ytEventSelection.cxx: Use PRWrandomRunNumber instead of using random_run_number
2. Change the comment style about skim part

v08. Modify the code for optimization.

1. LinkDef.h: Add DileptonTriggerWeight.h
2. ytEventSelection.h:
   • Comment yt_skim related part
   • Comment the weighted histograms
   • Comment the canvas
   • Add set_derivation_stat_weights()
3. yt_optimization.h: Arrange the order of codes.
4. run_bkg_samples.sh and run_signal_samples.sh: Add optimization in the script.
5. yt_useful_functions.cxx: Use SetPtEtaPhiM() for the muon and define mu mass
6. yt_cutflows.cxx: Comment pileupwgh
7. yt_optimization.cxx:
   • Uncomment h_derivation_stat_weights
   • Use GeV for pT, met, meff
8. cutflow.cxx:
   • Add isCutflow and isOptimization flag
   • Change submitDir based on the flag
   • Use merged_all_data.root
9. ytEventSelection.cxx:
   • Comment the skim related part, weighted histograms, derivation_stat_weights and the part to make plots
   • The Xsec relation calculationis are separated into signal and bkg. And correct the value if the sample doesn't be found.
   • The Nvtx, AvgMu, NLeps, Njets, mll histograms are recalculated using the correct weight.

v07. Add histograms of AvgMu and mll.

1. yt_cutflows.h: Add 3 methods:
   • get_mc_random_event_number()
   • get_AvgMu()
   • get_pileup_weight()
2. `ytEventSelection.h:
   • Add yt_skim
   • Add weighted and unweighted histograms of OSee and OSmumu AvgMu and mll.
   • Add new method set_isOptimization()
3. yt_cutflows.cxx:
   • Comment the argument PRW_weight and calculate the pileupwgh by myself.
   • Implement get_AvgMu() and get_pileup_weight()
4. ytEventSelection.cxx:
   • initialize(): put skim and optimization in the if condition
   • Add codes for declaring and filling AvgMu and mll histograms
   • Calculate random_run_number and pileup_weight
   • Correct the calculation of MC weight
   • Use random_run_number instead of using PRWrandomRunNumber

v06. Put cut 7 and cut 9 in if condition

1. Put cut 7 and cut 9 in if(!isSkim)
2. Add tag_pt_threshold in ytEventSelection

v05. Change histogram names

• Change histogram name from h_XXX_PRW to h_XXX_weighted
• histograms are separated into data and mc.

v04. Add two scripts

• Add run_bkg_samples.sh and run_signal_samples.sh to run over all samples.

v03. Add background samples

• The background samples are added in the cutflow.cxx

v02. Pass the correct value of DerivationStat_Weight

• If there are more than one root file in the dataset, the new way can pass the correct total DerivationStat_Weight to yt_optimization class.

v01. First commit.

The current version should work and was testing using v44 4topSM sample. But I have to run over all v44 bkg samples and compare with Ximo's results to make sure the results are correct.