Root

basic technique & information

Draw option

• same & histo hist->Draw("samehisto")

Fitting

• Fit option "N" --> Do not store the graphics function, do not draw
• If you want to get histogram, you should not use ->Eval(center value in a bin). The bin has a slope, so you should accommodate event weight in the bin. 1. GetMean(# of bin), you can get true center value. 2. f->GetRandom() => Fill it in a histogram.(When you get a value of f, you should set f->SetNpx(1000<).)

Disable "display TCanvas"

• root -b xxx.cxx
• gROOT->SetBatch();

DrawFrame()

• DrawFrame() is TH1 function, and then new TH1 object create and draw. If you'd like to add title(or xtitle, ytitle) in the canvas, you should set argument in DrawFrame(). c1-DrawFrame(xmin,ymin,xmax,ymax,"title;xtitle;ytitle")

Integral() vs GetEntries()

• GetEntries() show "all" entries.
• Integral() show summed entries in each bins (limited by max bin).

Disable Info message

• If you are annoyed by the output state coming from "SaveAs()" etc... , you can disable outputting those statements.
  #include "TError.h"
  gErrorIgnoreLevel = kWarning;
  

uncertainties

• Must do Sumw2() before Fill("",weight), if you store statistical uncertainties for each bin including in weight effect.

xAOD

• check containers and key names in xAOD

checkxAOD.py xAOD.pool.root

• check variables inside the containers

root xAOD.root
root [1] CollectionTree->Print("CotainerName*")

Analysis

Xsec

• Xsec list

For HQTTtResonanceTools

Boosted Jet Tagging

• We can select NLARGEJETTTRESSTTAU32_N or NLARGEJETTTRESSTMASS_N
• In default setting smooth cut (WP 80%) is applied.
• In this selection Lepton is required (Get the highest pT narrow jet "near" the lepton = selJet)
• "good" large jet have "pT > 300 GeV","|eta| < 2.0","std::fabs(top::deltaPhi(*largeJet, *lep)) > 2.3","top::deltaR(*largeJet, *selJet) > 1.5)"

How to get variables of ungroomed Jet

• AntiKt10LCTopoTrimmedPtFrac5SmallR20JetsAuxDyn.Parent can be used. (you can ignore "AuxDyn")
• AntiKt10LCTopoTrimmedPtFrac5SmallR20Jets : DataVector<xAOD::Jet_v1>
• you can get the object Ungroomed Jet.
  

const ElementLink<xAOD::JetContainer>& linkToUngroomed  =
     jetPtr->getAttribute<ElementLink<xAOD::JetContainer> >("Parent");
const xAOD::Jet* ungroomedJet1 = *linkToUngroomed;

• In the case "AntiKt10LCTopoJets" is Ungroomed Jet, so you should add "AntiKt10LCTopoJets" container in DxAOD.
  
• check # of ungroomed track jet. whether the variable is true or false.(2015.12.08)

optimization of selection

• There are some kinds of equation for sensitivity. I should choice the best equation for each situation.
• One of sensitivity equation is /sigma = S/(Q/2 + sqrt(B)). (S : Signal yields, B : Background yields, Q : Quality of sensitivity = signal intensity)
  The equation is stable even if B = 0;
• I should not use selection if the selection is at steep sloop. The errors at steep sloop are very big.
• Selected equation : σ = S_i/sqrt(S_i + B_i + (Δ*B_i)²)
  .(Δ=systematic uncertainty)

Lepton Veto in EXOT7

• electron: skimmed pass tight selection && pT>30GeV && |eta|<2.4 && # of electron >= 1
  
• muon : skimmed pass DFCommonGoodMuon selection && pT>30GeV && |eta|<2.4 && # of electron >= 1
  
• It seems to be small effect. (2016.01.15)

Event Selections

Preselection

• Trigger : HLT_ht_850
• Fat jet selection :
  • JetContainer : AntiKt10LCTopoTrimmedPtFrac5SmallR20Jets
  • mass > 100 GeV
  • pT > 200 GeV
  • |η| < 2.0
• track-bjet selection :
  • TrackJetContainer : AntiKt2PV0TrackJets
  • pT > 25 GeV
  • |η| < 2.5
  • MV2c20 70% WP(-0.3098) pass = "-0.3098 < MV2c20"
  • fatjet dR < 1.5 (if I count btagged jets in fatjet)
• ttbar reconstruction
  • # of fatjets == 2
  • # of b-jets >= 2
  • # of b-jets in (2nd)leading fat jet >= 1
  • tau32_wta < 0.71

Best Selection(provisional)

• For 2.5 TeV
  • leading fatjet mass < 210 GeV
  • 2nd leading fatjet mass < 190 GeV
  • leading fat jet track width < 0.7
  • 2nd leading fat jet track width < 0.4
  • # ghost tracks(leading )< 80
  • # ghost tracks(2nd leading )< 110
  • pt_balance < 0.6
  • sensitivity = 0.344802(sig : 11.8689, bg : 1083.88)
• For 1.5 TeV
  • leading fatjet mass < 210 GeV
  • 2nd leading fatjet mass < 190 GeV
  • leading fat jet track width < 0.6
  • 2nd leading fat jet track width < 0.6
  • # ghost tracks(leading )< 80
  • # ghost tracks(2nd leading )< 110
  • d_eta < 2
  • 2.5 < d_phi
  • pt_balance < 0.5
  • sensitivity = 5.03436(sig : 150.18, bg : 802.466)
• latest selection is not divide(2016.1.12)

Systematics and Statistic uncertainty

• In QCD_JZ3W events passed event selections are only "8", but # of events should be "O(100)" when events scaled for luminosity(3.34fb-1).
  As the result of this, the sample has large uncertainty.

For Data

• Trigger menu
• For Trigger : TRIGDEC HLT_ht850_L1J100 (TRIGDEC = trigger passed, TRIGMATCH =?)

Note

• sensitivity used # of ghost tracks is better than not used.
• # of tracks in ungroomed jet -> bat
• pbook command : retry(RUN#,retryBuild=True,newSite=True)

To do List

• Check the method of tau32 selection whether I should use pT dependency cuts or not.
• Check top::deltaPhi(*largeJet, *lep) and top::deltaR(*largeJet, *selJet)
• Check the procedure of HepTopTagger.

-- ShotaSuzuki - 2015-11-18