TWiki> Main Web>TWikiUsers>ShinShanYu>MasterElog2014 (2015-05-14, YuhsiangChang) EditAttachPDF

E-Log 2014/2015

Contents:

Yu-Hsiang Chang

2015/5

Show Hide

  • 2015/5/14 rivet produce 500k events with pythia6. data in di-muon, di-electron mass plot use pre-FSR detector acceptance, in combined channel mass plot uses pre-FSR full acceptance and in muon rapidity plots use pre-FSR detector acceptance.

  • di-muon mass plot.:
    d02-x01-y03.png

  • di-electron mass plot.:
    d03-x01-y03.png

  • combined channel mass plot.:
    d04-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 20 to 30.:
    d05-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 30 to 45.:
    d06-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 45 to 60.:
    d07-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 60 to 120.:
    d08-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 120 to 200.:
    d09-x01-y01.png

  • di-muon (absolute) rapidity plot in mass 200 to 1500.:
    d10-x01-y01.png

2015/3

Show Hide

  • 2015.3.23 chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.:
    chi2-2D-1.png

  • unnamed.pdf: chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.

  • unnamed-2.pdf: chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.

  • unnamed-3.pdf: chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.

  • unnamed-4.pdf: chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.

  • unnamed-5.pdf: chi2 in 2D plot. The X axis is the leptonic term, and the Y axis is the hadronic term.

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_with_incorrect_reset_histogram_.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_ttbar_three_component_pull1_and_pull2.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_ttbar_three_component_result_and_true.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_with_adding_tau_result_and_ture.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_with_adding_tau.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_ttbar_2_comp_and_Wbb_pull1_pull2.png

  • 2015.3.17 the result of incorrect setting of reset histogram. After correct the setting of reset, the true, the fit result of signal fraction, the pull(sig-true/sigerr), the fit-bias(pull2 or sig-true/true). The first 2 are fitted with 3 ttbar component. The following 2 are fitted with 2 ttbar component(adding tau component). The last 2 are fitted with 2 ttbar + 1 Wbb component.:
    fit_ttbar_2_comp_and_Wbb_result_and_true.png

  • 2015/3/11 compare the shapes in tau channel(red) and the shapes in muon channel(blue).:
    compare_tau_channel_signal.png
    compare_tau_channel_bk1.png

2015/2

Show Hide

  • 2015/2/24 Test the fit macro (TMinuit) with three component, the siganl and background1, background2, i.e 2 parameters. pseudo_data_Sig_0.3_bk1_0.6_bk2_0.1.png:
    pseudo_data_Sig_0.3_bk1_0.6_bk2_0.1.png fit_result_3_components.png:
    fit_result_3_components.png shut_screen_fit_result_3_components.png:
    shut_screen_fit_result_3_components.png

  • 2015/2/11 pseudo_data_SigFraction_0.7.png:
    pseudo_data_SigFraction_0.7.png

  • 2015/2/11 fit_for_fraction_0.7.png:
    fit_for_fraction_0.7.png

  • Use WJet sample without W pT cut, cut jet pT 200, then plot W-pT:
    Nocut_sample_gen_W_pT_after_cut_Jet_pT200.png

2015/1

Show Hide

  • 2015/1/12
    • Calculate MC sample corresponded luminosity, N-1 cuts study for gen semi-leptonic, signal and background shape with more statistics 2015_1_12.pdf.

2014/12

Show Hide

  • 2014/12/30
    • Correct rate with different B-tagging levels and the classification of ttbar events for tau21 efficiency measurement 2014_12_30.pdf.

2014/11

Show Hide

  • 2014/11/6
    • The study of deltaR(gen mu, gen nu) and deltaR(reco nu, reco mu) 2014.11.6.pdf

2014/10

Show Hide

  • 2014/10/30
    • The study of neutrino pz 2 sulutions. Which one is the correct one? Study the deltaR with reco Muon 2014.10.30.pdf.

  • 2014/10/23
    • Increase ttbar sample's statistics by command hadd. The lep_top_mass and had_top_mss of the combined 11 files (1 file has 5000 events) are shown as lep_top_11files.png and had_top_11files.png.
    • The next step is to study the gen-level particle to use the correct neutrino and b-jet and further reconstruct top more correctly.

  • 2014/10/16
    • write the macro to reconstruct top in leptonic channel and in hadronic channel 2014_10_16.pdf.

  • 2014/10/1
    • Attend the class of astrophysics.
    • Go to National Tsing Hua University for the collider physics class.

2014/9

Show Hide

  • 2014/09/30
    • Create and submit 3463 ttbar jobs to running.

  • 2014/09/29
    • Go to National Tsing Hua University for the elementary particle class .

  • 2014/09/25
    • Attend the class of astrophysics.
    • Using the lxplus batch system to compile for the ttbar production, and it works.
    • Together with the help of Henry, Ji-Ken, we fix the wireless and Chinese type setting of the MacBook Pro (laptop) and install the window 7 to my desktop.

  • 2014/09/24
    • Attend the class of astrophysics.
    • Go to National Tsing Hua University for the collider physics class.

  • 2014/09/23
    • compile for the ttbar production but fail.

  • 2014/09/22
    • Go to National Tsing Hua University for the elementary particle class .

  • 2014/09/19
    • compile for the ttbar production but fail.

  • 2014/09/18
    • Attend the class of astrophysics.

  • 2014/09/17
    • Attend the class of astrophysics.

  • 2014/09/15
    • Go to National Tsing Hua University for the elementary particle class.

  • 2014/09/12
    • Read and compare the JHEP09(2012)029 and PRD 84, 092004 (2011) to see the jet algorithm (AK5 or CA8) for the b-jet in ttbar semi-leptonic control sample, and have meeting with Eiko 2014_9_12.pdf.

  • 2014/09/10
    • Prepare the HEPDATA for our SMP-12-004.

  • 2014/09/09
    • Write macro for the correlation matrix of SMP-12-004 data.

  • 2014/09/05
    • Discuss code with Chih-Shen.
    • Read Roger Barlow's Statistics book for the correlation matrix of SMP-12-004 data.

  • 2014/09/04
    • Read HEPDATA website.
    • Discuss code with Chih-Shen.

  • 2014/09/01
    • Read how to measure the scale factor of 2 previous papers (CMS-PAS-EXO-11-006 and CMS-PAS-JME-13-006).
    • Read how to select ttbar event in muon control sample PhysRevD.84.092004

2014/8

Show Hide
  • 2014/08/28-29 *personal leave for 2 days.

  • 2014/08/26
    • Read the paper for how to select the ttbar control sample.
    • Discuss code with Chih-Shen.

  • 2014/08/25

  • 2014/08/20-22 *personal leave for 3 days.

  • 2014/08/19

  • 2014/08/14
    • personal leave

  • 2014/08/13
    • Read the AN for using ttbar events to measure the tau21 efficiency.
    • Discuss code with Chih-Shen.

  • 2014/08/12
    • Read the AN for using ttbar events to measure the tau21 efficiency.
    • Discuss code with Chih-Shen.

  • 2014/08/11
    • Test my grid certificate and discuss with Jun-Yi.
    • Discuss code with Chih-Shen.

  • 2014/08/08
    • Apply for the grid certificate (firstly the CMS VO).
    • Discuss code with Chih-Shen.
    • Read the AN for using ttbar events to measure the tau21 efficiency.

  • 2014/08/07
    • Compile and test the new code given by Eiko.
    • Discuss code with Chih-Shen.

  • 2014/08/05
    • read the 3 ANs
    • Help and discuss with Jun-Yi and Chih-Shen for code and root problem

Jun-Yi Wu

Personal github repo: https://github.com/juwugit/XtoZh

2015/03

Show Hide
  • 03/13
    • 20150313daily_pythia_vs_herwigpp.pdf: 20150313daily_pythia_vs_herwigpp.pdf
    • Result of comparison between PYTHIA8 and Herwig++
    • Note that, the result shows Herwig++ has narrower mLL distribution and higher peak on both Z/jet Pt.
    • LO Herwig++ sample shows discripency on high JetMultiplicity region.
    • Rapidity difference plot shows large discripency between PYTHIA8 and Herwig++.

  • 03/05
    • 20150305daily_juwu.pdf: 20150305daily_juwu.pdf
    • fix the wrong sb&sig region definition, it should be distinguish by the prunedmass of leading jet, not prunedmass of every jets.

  • 03/03
    • 20150303daily_juwu.pdf: 20150303daily_juwu.pdf
    • Slide for CSV value of fatjet/subjet and their alpha ratio, setting higher jet Pt cut(200GeV) and sideband region shifting to 70GeV~110GeV

2015/02

Show Hide

2015/01

Show Hide
  • 01/12

  • 01/06
    • Spin-0 high priority LHE files are produced, stored in pcncu:/data4/juwu/XtoZh_new_LHE
    • Fitted histogram of radion mass here: http://imgur.com/oUrgfjq
    • For spin-2, one channel haven't finished, still running in LSF system. It's RSG_WW_WhadWhad, three mass point left.
    • Fitted histogram of RSG mass here: http://imgur.com/wCvvmQL

2014/12

Show Hide
  • 12/29
    • All new Spin-1 signal LHE files are successfully completed and stored in pcncu:/data4/juwu/XtoZh_new_LHE (except Zprime_Zh_Zlephbb, it's already produced)
    • 20141229juwu_daily.pdf: 20141229juwu_daily.pdf

  • 12/09

2014/11

Show Hide
  • 11/25
    • h5_xzhqq.png:
      h5_xzhqq_all_MG5.1.5.14.png

  • 11/24
    • h_xzhqq.png:
      h_xzhqq.png

  • 11/20
    • btag_M1500_new.png:
      btag_M1500_new.png
    • this figure already change the fatjet definition. The numerator is the events pass the CSV cut and other all selections while the denominator is events pass all selections.

  • 11/19
    • btag_M1500.png:
      btag_M1500.png

  • 11/10
    • After set asrwgtflavor(highest quark flavor for a_s reweight) from 5 to 4 in two samlpes(dy01j_LO and dy0j_LO)
    • Rapidity diff. and sum:
      canvas6.png
    • Jet kinematics and multiplicity:
      canvas5.png
    • Z kinematics and dilepton mass:
      canvas4.png
    • 20140929juwu-Validation_ZplusJets.pdf: my report on 09/29 matrix element meeting

  • 11/05
    • subjet CSVL: eff:0.232176 mistag:0.0104701
    • subjet CSVM: eff:0.117916 mistag:0
    • subjet CSVT: eff:0.0511883 mistag:0
    • fatjet CSVL: eff:0.700906 mistag:0.145927
    • fatjet CSVM: eff:0.483384 mistag:0
    • fatjet CSVT: eff:0.239275 mistag:0

2014/10

Show Hide

  • 10/15
    • update result: comparison between data and MC, using ABC data sets and plot jetId variables, tau21, prunedjetmass, b-tagging variables.
    • https://github.com/syuvivida/xtozh_common/tree/master/JunYi/JetID_study
    • WZ flattutple is done, located in pcncu data1
    • Now copying ZZ flattuple from TW tier2, it costs one day.
    • Last hadronizer for dy01j_LO is running.
    • Resubmit bad jobs of DY100 MC background.

2014/09

Show Hide

  • /09/29
    • WW/WZ/ZZ: Resubmit bad jobs to CRAB, almost done, about 70 jobs remain for each sample.
    • Study CMS PAS BTV-13-001 for b-tag algorithm.

  • 09/26
    • Prepare slide for generator meeting.
    • Separate the two lepton channels when plotting cut efficiency: ee channel ; mumu channel

  • 09/25
    • Efficiency of various cuts: figure link
    • For the denominator, it is the total entries that adds previous cut, ex: denom in 1st bin is the total entries, denom for 2nd bin is the number of entries after lepton ID selection.

  • 09/23
    • finish background estimate for electron channel data(A+B): gitlink, muon data is lack of statistics (only A).

  • 09/19
    • Found a mistake on running hadronizer for 5f NLO, re-run it again.
    • Reconstruct X particle mass for sideband, signal region and calculate alpha ratio. electron channel is done.
    • Submit jobs through CRAB2 on SLC6, 1-500 first submit for WW/WZ/ZZ.

  • 09/18
    • 4 flavor results of aMC@NLO gridpack validation: aMC@NLO plot (canvas1~3 for 5f , 4~6 for 4f). you may see the same trend of distribution between 4f and 5f.
    • test CRAB2, and it's fine to use it on SLC6, already submit 10 jobs for testing WW sample.

  • 09/17
    • Add Higgs Pt (CA8jet Pt) and prunedjet mass to MC-data-comparison. data_MC_comparison
    • parton shower for 4f lhe files finished, start to make some plot after I copy the ntuples to my laptop.
    • Read X->ZZ analysis note and Andrea's thesis.
    • test to use crab on SLC6.

  • 09/16
    • Add ratio to these two variables data_MC_comparison
    • Resubmit failed jobs(WW/WZ/ZZ) again.
    • One left lhe file to finish PYTHIA8 (dy0j_4f_LO_MLM.lhe), herwig++ testing is okay(for dy0j_4f_LO_MLM.lhe).

  • 09/12
    • WW/WZ/ZZ jobs still about 2/3 jobs failed, add one server to blacklist and resubmit them.
    • 3 drell-yan samples finish parton shower and transform to ntuple.
    • do some simple data,MC comparison. it is mLL and tau21. data_MC_comparison

  • 09/11
    • WW/WZ/ZZ jobs finished, start to rerun failed jobs.
    • send pythia8 parton shower jobs(dy0j_LO, dy0j_NLO, dy01j_LO) to lxplus batch

  • 09/06
    • Update cards for gridpack dy_gridpack_cards
    • Submit last 3 jobs (2001-2003) of WZ sample, 1501-2000 jobs of WW and 1501-1965 jobs of ZZ. Note that, a lots jobs which are submitted last time failed (about 30 jobs of WZ and ZZ, two hundred jobs of WW failed), get wrapper exit codes.

  • 09/05
    • Cards for gridpack production put in here dy_gridpack_cards.
    • JetSelection.h latest version, note that tau21 cut hasn't been applied yet until I finish optimization.
    • Submit 1501-2000 WZ jobs and 1001-1500 WW/ZZ jobs to crab.

  • 09/03
    • Submit 1001-1500 WZ jobs and 501-1000 WW/ZZ jobs to crab.
    • Drell-yan plus 0 jet LHE file finished showering(PYTHIA8). Start to use delpanj package to analyze GEN.root.
    • Read the latest email for the models you have to use when generating 5-flavor production

  • 09/02
    • Update JetSelection.h
    • LHE file generation using gridpack succeed(dy+0jet 5 flavor 100k events), start to send pythia8 parton shower job to batch system.
    • Start to submit 501-1000 jobs to crab for WZ sample. (WW and ZZ still have unfinish jobs)

  • 09/01
    • Test gridpack interface, tarball generation is fine, but when I start to generate events, it fails.
      • Comment: Read the email I sent to you and Hien on August 26 and you will find the solution in item 5.
    • About pat-tuple production, I re-define the host name in blacklist, hope this will be working
    • Submit first 500 jobs of WW, WZ, ZZ sample

2014/08

Show Hide

  • 08/26
    • JetSelections.h now does sorting using std::map function. Check update on github. JetSelection.h
    • New Tau21 analyzer using Henry's lepton selection header file is completed, but I need to download new ntuples(it needs IsoTrack infos.)
    • NCU tier3 is back, start to re-submit first 500 jobs to CRAB. (WW)

  • 08/14
    • Add ttbar to background, tau31 tau41 to analyzer.

  • 08/13
    • Re-make significance figure using new drell-yan background. See updated figures here https://juwu.web.cern.ch/juwu/XtoZH_plots/Tau21/
    • The result of optimization cut doesn't change.
    • Question from Eiko: did you use the cross section and number of processed events from the twiki page ? What is the status of WW/WZ/ZZ patTuple production

  • 08/12
    • When I calculated scale factor for ttbar background sample, I found that drell-yan scale factor is wrong, and it already fixed.
    • Add ttbar sample to background, plot tau21 significance again. (not finish yet)
    • 20140812juwu-NCU_daily_eleID.pdf: add sigmaIetaIeta

  • 08/10
    • Read analysis note AN2014_063_v6 and AN2013_040_v7 , also the thesis of Andrea(Italian student of U. PISA)

  • 08/09
    • Test crab job succeed, starting to submit 3 MC background samples(WW,WZ,ZZ) to CRAB.
    • Successfully compiled ntuple producer package, begin to submit first job, WW sample.

  • 08/06
    • Try to separate two lepton channels and then remove overlapping, the peak near 0.1 reduced but still there. Use only leading lepton after separate two channels to calculate deltaR and remove overlapping, it works. The peak is gone. Maybe I should do this when I first met this issue, fake jet almost come from leading leptons.
    • Plot Tau21 of llqq_M1500 signal sample, plot signal and background efficiency on the same canvas, plot tau21 significance. https://juwu.web.cern.ch/juwu/XtoZH_plots/Tau21/
    • Note that, both signal and background samples are normailzed before calculating significance. And both samples pass Xmass window cut, which I didn't apply this cut to old ntuples when I was studing tau21.

  • 08/05
    • Edit the analyzer code, make it simpler.
    • Derive errors of efficiency.
    • Discuss some root problem with Yu-Hsiang and undergraduate student.

  • 08/04
    • Remove overlap between jet-lepton in my analyzer code, but there's still a peak near 0.1.
    • Discuss nsubjettiness with Yu-Hsiang.

Henry Tong

2014/12

Show Hide

  • 2014/12/05
    • This is the summary slide of 13TeV boosted muon
The summary slides of 13TeV muon

2014/11

Show Hide

  • 2014/11/20
    • I plot the the distributions of the invariant mass of Zprime by using ttbar + all diboson samples. Below is the result:
bkgZpMass

  • 2014/11/19
    • The alpha ratio fit is done.
      • Chi-Square = 9.98926
      • Degree of freedom = 13
      • Probability of fit = 0.694812
    • The Chi-Square is calculate from: Sum of( (red-curve-by-const-bin - alpha-histo-by-variable-bin)^2 / (error-of-alpha-histo-by-variable-bin)^2 )

variable-bin-width fix-bin-width

2014/10

Show Hide

  • 2014/10/28
    • The event efficiency of dileptons are been plotted. But I am not sure the algorithm of electron efficiency are correct or not. The link below is the code, and also the efficiency of dimuon and dielectron.
ele code dimuon efficiency dielectron efficiency

  • 2014/10/07-2014/07/08
    • Muon and electron Data/MC results has been updated, as link below:
    • The variables are been plot according to so called n-1 cut, which means that when you draw a specific variable of lepton, you don't do any cut on that variable but cut on the other variables. Then, we plot the Data/MC ratio in order to show the agreement of MC and data.
    • Why n-1? It is because we need to check the agreement of that variable, so we need to keep its original information, so dont apply cut on it.
    • Why cut? It is because we need to check the agreement of variable which is contribute by a good leptons (leading pt and sub-leading pt lepton))
    • Why in some plot, the ratio on some region far from one? This is because of many reason, one is the event at that region is not enough, the other is the MC cannot simulate well that region.
    • From the above question, but why in the variable plot, the data and MC looks agree fine? This is because the variable plot shows absolute error, but what we care about is actually relative error.
muon variable electron variable

2014/09

Show Hide

  • 2014/09/30
    • Today I finish redraw the lepton variables with adding a plot of data/MC ratio.

  • 2014/09/03-2014/09/07
    • Try to write my report, and submit the crab job for double_mu_C and D. The C jobs are almost done, left not more than 100 resubmited jobs; the D jobs are still running, now is 501-1000.

  • 2014/09/02
    • I plan to spend one week to write my NSC report.

2014/08

Show Hide

  • 2014/08/28
    • Today I plot the comparison of lepton ID variables with DY backgrounds.
    • Below is the plot of muon variable. The variables are been plot after passing the passMuonID.C header file.
    • Below are the plots of electron variable. The variables are been plot by using n-1 method, which is when you plot a variable x, don't cut the x but cut the other variables.

  • 2014/08/27
    • Today I plot the stack diagram of background pruned jet mass distribution, compare with data. The results: backgroundStack.gif
    • The trigger cut has been applied. The total event of signal didn't reduce. Our lepton selections are more tightly compare with the trigger cut.

  • 2014/08/26
    • Today I combine the two DY+Jet backgrounds, and plot the alpha ratio, which is the ratio of the signal region and the side band region of reconstructed Z prime mass distribution, as a function of Zprime mass.
    • The distribution are plot with not uniform width bins, acroding to the decreasing statistic in the high mass tail.
    • The alpha ratio allows a prediction of the mass spectrum in the signal region starting from the measured distribution in the side band.
    • The link of my results: backgroundEstimation.gif
      • Total event of bkg side band region : 13.4858
      • Total event of bkg signal region : 3.09481
      • Total event of data side band region: 8
      • Total event of data signal region : 3

  • 2014/08/25
    • MCFM's resuts: (central value only) (scale = -1, facscale = -1)
Process M_34 Min_56 Central value(fb)
      removebr = true, zerowidth = true removebr = false, zerowidth = false
      part = tota part = lord part = tota
W+ $\rightarrow$ ($\nu$ $\mu$+) + Z0 $\rightarrow$ ( e- e+ ) 0 12 14355.564 9016.412  
W+ $\rightarrow$ ($\nu$ $\mu$+) + Z0 $\rightarrow$ ( 3 x ($\nu$ $\nu$) ) 0 12 309.149 194.174 304.166
W- $\rightarrow$ ($\nu$ $\mu$-) + Z0 $\rightarrow$ ( e- e+ ) 0 12 8079.267 5034.071  
W- $\rightarrow$ ($\nu$ e-) + Z0 $\rightarrow$ ( 3 x ($\nu$ $\nu$) ) 0 12 173.923 108.371 170.935
Z0 $\rightarrow$ ( e- e+) + Z0 $\rightarrow$ ( $\mu$- $\mu$+) [note] 12 12 7773.508 5554.789  
    • [note] : If part = lord, removebr = true, zerowidth = false, Central value = 13157.912 fb.

  • 2014/08/20
    • This is the table of comparison of electron's ID cut value.

Variables Barrel End cap
  Andrea's cut myCut Andrea's cut myCut
eleDelEtaIn (<) 0.005 0.004 0.007 0.005
eleDelPhiIn (<) 0.06 0.03 0.06 0.02
eleSigIhIh (<) no cut value 0.01 0.03 0.03
eleHoE (<) 0.05 0.12 0.05 0.1

  • 2014/08/18
    • Today I check the ID efficiency of electrons. I found that if I comment the cut of eleDelEtaIn, eleDelPhiIn, eleSigIhIh, the ID efficiency will be so flat.

  • 2014/08/13
    • To optimize the signal and background today. I try to add the three backgrounds and draw the iso/pt distribution of it.

  • 2014/08/12
    • Today I reconstruct the mass of Z prime, in order to choose a better mass window cut. The conclusion is, we choose the Z prime mass +- 12% as our mass window.

  • 2014/08/11
    • The muon isolation cut efficiency in reconstructed level has a drop at low delta R region. This maybe cause by the inefficiency of detector for low delta R case. I plot the track isolation at generator level, the efficiency are almost at 100%. Then, I plot the pt ratio of (muon of fail the ID cuts and isolation cuts) with its generator level muon, realize that the reconstructed pt are much more larger or lower then the generator level pt, cause the isolation cut are incomplete. I will process to the optimization of background today.

  • 2014/08/08
    • Continue study the isolation cut efficiency. I try to change the symmetrical uncertainty to the asymmetrical uncertainty.
    • symmetrical uncertainty: Useless or lack of information when the uncertainty is 100% or 0. We don't know how much events contribute the efficiency, and we cannot get the weight when we need to compare or fitting.
    • asymmetrical uncertainty: When the event distribution is not same as Gaussian distribution (a symmetry distribution)(maybe a lot of events at small region, or the reason above), the RMS cannot equal to sigma, so we need an asymmetrical uncertainty to describe well how these event locate.

  • 2014/08/07
    • After the CorrIso cut, I found that the efficiency is much lower then expected at low delta R region, so I try to write an isolation cut algorithm by myself to check the reason.

  • 2014/08/06
    • eta: "angle" of (direction of momentum at initial point) with Z-axis. It is physics eta.
    • super cluster eta: "angle" of (the line of center of collide and the detected point at detector ) with Z-axis.

  • 2014/08/05
    • To check why the ID efficiency as a function of pt will decrease slightly at higher pt region for the sample of hz->qqll at M_X=2000.

  • 2014/08/04
    • ID efficiency is an efficiency of signal, when we pass the selections to the data, it can throw away some background but also some signal, so we need this efficiency to check how much signal we lose.
    • The reason of studying the efficiency as a function of variables is that the pt and eta distribution are not same for many processes. We use Monte Carlo generated process to check the efficiency of process which has been study long time, and compare with the same process comes from control sample, to find the weight of the efficiency. So this weight can applied to our ID efficiency of new process which is generated by Monte Carlo. After this, the efficiency can be approximate to the real data.
    • To check why the ID efficiency as a function of pt decrease at low region. The reason is because some secondary leptons which have low efficiency lower the total efficiency. These are the backgrounds come from jet events or other process.

Chao-Wen Huang

2015/04

Show Hide

  • 2015/04/16

-- ShinShanYu - 04 Aug 2014

Attachments Attachments
Topic attachments
I Attachment History Action Size Date Who Comment
PDFpdf 2014.10.30.pdf r1 manage 194.3 K 2014-10-30 - 09:43 YuhsiangChang  
PDFpdf 2014.11.20.pdf r1 manage 346.0 K 2014-11-20 - 09:25 YuhsiangChang  
PDFpdf 2014.11.6.pdf r1 manage 300.4 K 2014-11-06 - 09:52 YuhsiangChang  
PDFpdf 2014_10_16.pdf r1 manage 226.0 K 2014-10-16 - 11:08 YuhsiangChang  
PDFpdf 2014_10_21_Yuhsiang.pdf r1 manage 250.3 K 2014-10-21 - 10:28 YuhsiangChang  
PDFpdf 2014_12_30.pdf r1 manage 50.5 K 2014-12-31 - 05:13 YuhsiangChang  
PDFpdf 2014_8_15_data_MC_scale_fator_of_tau21_efficiency_in_ttbar_sample.pdf r1 manage 571.6 K 2014-08-15 - 10:09 YuhsiangChang  
PDFpdf 2014_9_12.pdf r1 manage 586.3 K 2014-09-12 - 09:29 YuhsiangChang  
PDFpdf 2014_9_2_measure_scale_factor_of_tau21_efficiency_in_ttbar_control_sample.pdf r1 manage 1059.9 K 2014-09-02 - 09:45 YuhsiangChang  
PDFpdf 2015_1_6.pdf r1 manage 395.3 K 2015-01-06 - 09:32 YuhsiangChang  
PDFpdf AN2013_045_v7.pdf r1 manage 7914.6 K 2014-08-06 - 08:56 YuhsiangChang 3 ANs that use tau21 to discriminate signal and background in the case of boosted jet
PDFpdf AN2013_139_v11.pdf r1 manage 6063.8 K 2014-08-06 - 08:56 YuhsiangChang 3 ANs that use tau21 to discriminate signal and background in the case of boosted jet
PDFpdf An-Li_qualify_exam.pdf r1 manage 12115.5 K 2014-09-04 - 04:53 YuhsiangChang  
PDFpdf Gen_particle_ID.pdf r1 manage 168.0 K 2014-10-08 - 08:58 YuhsiangChang  
PDFpdf Group_Meeting.pdf r1 manage 75.3 K 2014-12-29 - 09:32 UnknownUser  
Texttxt HEPDATA_input_v2.txt r1 manage 7.6 K 2014-09-14 - 10:26 YuhsiangChang  
PDFpdf How_to_select_the_ttbar_control_sample.pdf r2 r1 manage 828.8 K 2014-08-27 - 16:56 YuhsiangChang  
PDFpdf Particle_Physics_3rd_Ed_2008_B.R.Martin_G.Shaw.pdf r1 manage 25207.0 K 2014-11-05 - 19:55 YuhsiangChang 11/5 the class collider physics report
PDFpdf TOP-12-020-paper-v10.pdf r1 manage 582.9 K 2014-08-19 - 10:35 YuhsiangChang CMS paper review
PNGpng ca8jetpt.png r1 manage 23.1 K 2014-10-08 - 10:32 YuhsiangChang Yu-hsiang ttbar Oct 8
PDFpdf collider_physics_report_part1.pdf r1 manage 54577.5 K 2014-11-05 - 19:54 YuhsiangChang 11/5 the class collider physics report
Unknown file formatpptx collider_physics_report_part1.pptx r1 manage 49588.1 K 2014-11-05 - 19:54 YuhsiangChang 11/5 the class collider physics report
Unknown file formatext correlation_matrix_of_PhotonJet r1 manage 6.4 K 2014-09-14 - 10:26 YuhsiangChang  
Unknown file formatext correlation_matrix_of_ZJet r1 manage 9.7 K 2014-09-14 - 10:26 YuhsiangChang  
PNGpng deltaR_genNu_recoNu_of_2pz_sol.png r1 manage 26.8 K 2014-10-28 - 09:29 YuhsiangChang  
PNGpng elept.png r1 manage 22.6 K 2014-10-08 - 10:32 YuhsiangChang Yu-hsiang ttbar Oct 8
PNGpng fix_jet_loop_result.png r1 manage 394.4 K 2014-12-16 - 09:21 YuhsiangChang  
PDFpdf had1.pdf r1 manage 14.2 K 2014-12-11 - 09:34 YuhsiangChang  
PNGpng had_top_11files.png r1 manage 21.8 K 2014-10-23 - 10:03 YuhsiangChang  
PDFpdf institutional_review_of_Top-12-020.pdf r1 manage 114.8 K 2014-08-19 - 10:35 YuhsiangChang CMS paper review
PDFpdf lep1.pdf r1 manage 14.1 K 2014-12-11 - 09:34 YuhsiangChang  
PNGpng lep_top_11files.png r1 manage 23.4 K 2014-10-23 - 10:03 YuhsiangChang  
PNGpng lep_top_after_chi2.png r1 manage 26.2 K 2014-11-18 - 09:33 YuhsiangChang  
PNGpng lep_top_before_chi2.png r1 manage 25.5 K 2014-11-18 - 09:33 YuhsiangChang  
PNGpng match_had_b.png r1 manage 23.2 K 2014-12-22 - 08:14 YuhsiangChang  
PNGpng match_had_w.png r1 manage 24.0 K 2014-12-22 - 08:14 YuhsiangChang  
PNGpng match_lep_b.png r1 manage 24.9 K 2014-12-22 - 08:14 YuhsiangChang  
PNGpng match_lep_nu.png r1 manage 24.7 K 2014-12-22 - 08:14 YuhsiangChang  
PNGpng match_rate_12_22.png r1 manage 410.1 K 2014-12-22 - 08:14 YuhsiangChang  
PNGpng mupt.png r1 manage 21.7 K 2014-10-08 - 10:32 YuhsiangChang Yu-hsiang ttbar Oct 8
PNGpng p1.png r1 manage 567.5 K 2014-12-22 - 09:27 YuhsiangChang  
PNGpng p2.png r1 manage 569.5 K 2014-12-22 - 09:27 YuhsiangChang  
PNGpng p3.png r1 manage 572.5 K 2014-12-22 - 09:27 YuhsiangChang  
PNGpng p4.png r1 manage 568.9 K 2014-12-22 - 09:27 YuhsiangChang  
PDFpdf particleid.pdf r1 manage 13.8 K 2014-10-07 - 10:30 YuhsiangChang  
PNGpng pfMetCorrPt.png r1 manage 23.4 K 2014-10-08 - 10:32 YuhsiangChang Yu-hsiang ttbar Oct 8
PNGpng picture1.png r1 manage 576.3 K 2014-12-11 - 09:38 YuhsiangChang  
PDFpdf ref_34.pdf r1 manage 875.3 K 2014-08-27 - 16:56 YuhsiangChang  
PDFpdf ref_40.pdf r1 manage 1093.4 K 2014-08-27 - 16:56 YuhsiangChang  
PDFpdf ref_46_the_measure_ttbar_crosssection_in_leption_jet_channel_use_b_tagging.pdf r1 manage 783.2 K 2014-09-01 - 15:45 YuhsiangChang  
PNGpng study_match_had_b.png r1 manage 21.2 K 2014-12-22 - 09:27 YuhsiangChang  
PDFpdf tau21_and_AN-2013_-40_AN-2013_045_AN-2013_139_on_2014_8_6.pdf r1 manage 392.7 K 2014-08-06 - 08:49 YuhsiangChang 2014.8.6 tau21 with 3 ANs
PDFpdf top_tagging_algo_ref_27.pdf r1 manage 585.9 K 2014-08-27 - 16:56 YuhsiangChang  
Edit | Attach | Watch | Print version | History: r193 < r192 < r191 < r190 < r189 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r193 - 2015-05-14 - YuhsiangChang
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback