Higgs to Mu Mu Analysis @ UF

Code Overview and Github Repos

• Stage 1: https://github.com/acarnes/UfHMuMuCode
  • Runs over the raw samples and puts the info into a ttree
  • The master branch works with CMSSW_7_4_X
  • The CMSSW_7_6_X branch works with 7_6_X
• Stage 2: https://github.com/bregnery/UFhmumuVariableStage
  • Adds some higher level variables to the stage 1 information
• Analysis Code: https://github.com/acarnes/UFDimuAnalysis
  • Takes info from the stage1 samples to do different things for the analysis
  • Synchronize with Adrian
  • Makes N-1 plots and uses these to optimize the muon selection cuts according to Asimov significance

Samples

Data, Drell Yan, ttbar, VBF, GGF

Run through stage1 and available on the ihepa servers at /cms/data/store/user/t2/users/acarnes/h2mumu/samples/stage1

ttHmumu

Introduction

This is a summary of our private ttHmumu sample generation. Our current ttHmumu sample is accessible on the UF Tier 2 cluster, the UF HPC, and CERN's lxplus cluster.

Show ttHmumu Sample Location Hide ttHmumu Sample Location

Sample Generation

The most up-to-date version of the ttHmumu generator can be found at: https://github.com/bregnery/ttHmumuGenerator

Generation Process

• First, we created a ttHmumu cfg.py using a cfi.py file from Adrian and the following cmsDriver settings:
  • cmsDriver.py Configuration/Generator/python/h2mu_tth_M125GeV_13TeV_cfi.py \ --step GEN,SIM,DIGI,L1,DIGI2RAW,HLT:GRun,RAW2DIGI,L1Reco,RECO \ --pileup_input dbs:/MinBias_TuneA2MB_13TeV-pythia8/Fall13-POSTLS162_V1-v1/GEN-SIM \ --pileup AVE_20_BX_25ns \ --conditions PHYS14_25_V1 \ --customise SLHCUpgradeSimulations/Configuration/postLS1Customs.customisePostLS1 \ --magField 38T_PostLS1 \ --geometry Extended2015 \ --beamspot Realistic8TeVCollision \ --datatier GEN-SIM \ --mc \ --eventcontent AODSIM \ --python_filename h2mu_tth_M125GeV_13TeV_FULLSIM_cfg.py \ -n 10 --no_exec \ --fileout h2mu_tth_M125GeV_13TeV_pythia6_AODSIM_1.root
• Then, we updated the global tag in this cfg.py file to "MCRUN2_74_V9"
• Next, we used this file to generate the ttHmumu events
• Then, we converted all of the events to miniAOD format using a program provided by Justin: https://github.com/bregnery/ttHmumuGenerator/blob/master/convertMC_MINIAOD.py
• Finally, we used Justin's UFDiMuonsAnalyzer to create stage 1 ntuples: https://github.com/jhugon/UfHMuMuCode/tree/CMSSW709_miniAOD

Current Sample

Our current ttHmumu sample contains 98,700 events and has a constant pile up. Here are some plots summarizing our current sample (please note that these plots do not include an ET or Eta threshold):

Show ttHmumu Generation Plots Hide ttHmumu Generation Plots

Updates in progress

7-Aug-2015 the issue with Pile up distribution has been solved

Currently, there are no known issues with the ttHmumu sample.

vHmumu High Statistics Sample

Sample Generation

The most up-to-date version of the vHmumu generator can be found at: https://github.com/bregnery/vHmumuGenerator

Issues and Updates

31-May-2015 the .root files generated still need to be tested.

Code

New Variable Creation

We are creating a new program that adds new variables to the Stage 1 ntuples created by UF's analyzer (https://github.com/acarnes/UfHMuMuCode). This program is accessible at: https://github.com/bregnery/UFhmumuVariableStage

Current Program

Our current program creates the following:

• A new branch called "diJet" that includes leaves for mass, phi, eta, pt, and Et
• A new leaf called "VHphi" which is equivalent to diJet.phi - diMuon.phi
• A new leaf called "phiStar" for the variable phi star
• A new leaf called "Zeppenfeld" for the Zeppenfeld variable

Phi Star

Here is the section of our program that calculates phi star:

Show Phi Star Calculation Hide Phi Star Calculation

float muonDelPhi = TMath::Abs(reco1.phi - reco2.phi);  // reco 1 and 2 are the reconstructed muons
if(muonDelPhi > TMath::Pi()){
    muonDelPhi = 2*TMath::Pi() - muonDelPhi;
}
float phiACOP = TMath::Pi() - muonDelPhi;
float thetaStarEta = TMath::ACos(TMath::TanH((reco1.eta - reco2.eta)/2));
phiStar = TMath::Tan(phiACOP/2)*TMath::Sin(thetaStarEta);

Zeppenfeld Variable

Here is the section of our program that calculates the Zeppenfeld Variable:

Show Zeppenfeld Variable Calculation Hide Zeppenfeld Variable Calculation

if(jets.size() >= 2){
    double meanEta = (jets[index1].Eta()+jets[index2].Eta())/2;
    Zeppenfeld1 = jets[index1].Eta()-meanEta;
    Zeppenfeld2 = jets[index2].Eta()-meanEta;
}

New Variable Example Plots

These plots were created from the VHmumu stage 1 ntuple:

Show New Variable Plots Hide New Variable Plots

Plots, Results, and Presentations

Cuts

• The muon selection cuts used in run1
  • selection_cuts_info.pdf
• Optimizing the muon and event selections according to Asimov significance for run2
  • For Run2C &D 2.17 fb-1
    • nm1_6-1-16.pdf
  • For Run2C &D 20 fb-1
    • link_to_pdf

Improving the Mass Resolution

• Put Zach's talks here
• Need to find them from indico

Data/MC Agreement

• Comparing Data/MC for Run2C &D 2.11 fb-1 using the recommended minbias xsec value for pileupCalc
  • dy_vs_data_25ns_dec9.pdf
• Looking at different minbias xsec values to match the NPV distributions better
  • dy_vs_data_25ns_dec15.pdf

Useful Links and Papers

• H2mu Twiki
  • https://twiki.cern.ch/twiki/bin/viewauth/CMS/HiggsToMuMu
• Run1 Paper
  • http://www.sciencedirect.com/science/article/pii/S0370269315002117

Legacy Run 1 info

Miscellaneous

-- SergeiVGleyzer - 2015-08-04