qqH, H lepton jet CMSSW vs ORCA

1
qqH, H ? ?? ? lepton jetCMSSW vs ORCA
2
Software and Data Sample

• Compare ORCA (as of PTDRII study) and CMSSW
• Software
• ORCA_8_7_3 Jets package to remove fake jets
  around ? 3
• (equivalent to ORCA_8_7_4)
• CMSSW_1_3_4
• Dataset (use MH125GeV/c2)
• ORCA hg05t_qqh_2tau_lj_125 (L 2x1033
  cm-2s-1)
• 50k events (qqH, H ? ?? ? lepton jet)
• CMSSW /qqH125_2tau/CMSSW_1_3_1-Spring07-1137/GEN-
  SIM-DIGI-RECO
• (no pileup)
• 96k available, 20k processed (all tau decay ?
  select only leptonjet 9k)
• Main difference in the reconstruction stream
• In ORCA, start with lepton search and then taujet
• in CMSSW, start with taujet search and then lepton

3
Object Reconstruction and Selection I
3
Imperial College London
4
Object Reconstruction and Selection II
4
Imperial College London
5
Event Selection

• Triggers skipped for now
• pT (ET) cuts acceptance on lepton, ?-jet and
  forward jets,
• pTlep gt 15GeV/c, (? lt 2.5)
• pT?-jet gt 30GeV, ? lt 2.3
• pTjet gt 40GeV , ? lt 4.5
• Other background rejection cuts
• ??j1j2 4.2
• ?fj1j2 2.2
• M(j1,j2) gt 1TeV
• MT (lepton,MET) lt 40GeV
• (Central jet veto is not implemented in CMSSW
  yet)

6
Distribution of Physics Objects(Generated and
Reconstructed)For reconstructed objects, only a
simple ID cuts applied and not the pT (except
for jets there is a minimum pT 20 GeV)
7
Kinematics (Gen) Higgs
ORCA CMSSW
Y-axis of events
7
Imperial College London
8
Kinematics (Gen) Tau Leptons
ORCA CMSSW
Y-axis of events
8
Imperial College London
9
Kinematics (Gen) Leptons
ORCA CMSSW
Y-axis of events
9
Imperial College London
10
Kinematics (Rec) Leptons
ORCA CMSSW
Y-axis of events
10
Imperial College London
11
Kinematics (Gen) TauJets
ORCA CMSSW
Y-axis of events
11
Imperial College London
12
Kinematics (Rec) TauJets
wider acceptance for CMSSW
ORCA CMSSW
Y-axis of events
12
Imperial College London
13
Kinematics (Gen) Quarks (parton)
ORCA CMSSW
Y-axis of events
13
Imperial College London
14
Kinematics (Rec) Tagging Jets
ORCA CMSSW
Y-axis of events
14
Imperial College London
15
Kinematics (Gen Rec) MET
solid calo muon dashed type1 filled
high pT obj
ORCA CMSSW
Y-axis of events
15
Imperial College London
16
Kinematics (Gen) VBF parameters etc
ORCA CMSSW
to be investigated
Y-axis of events
16
Imperial College London
17
Kinematics (Rec) VBF parameters etc
ORCA CMSSW
Y-axis of events
17
Imperial College London
18
Distribution of Basic ObjectsTracks only
reconstructed saved in tree at the momentJets
genjets recjets
19
Tracks and Jets Multiplicity
gen jets
rec tracks
ORCA CMSSW

• Multiplicity higher for CMSSW despite not having
  pile-up!
• rec tracks satisfying cuts on page 3
• ? could be that the N hits requirement is too
  low in CMSSW
• gen jets all hadron jets with etalt4.5
  pTgt20GeV
• ? I cant think of an obvious reason here
• More plots to follow on this

19
Imperial College London
20
Kinematics (Rec) Tracks
looks like some fakes from low N hits
ORCA CMSSW
Y-axis N tracks / event
20
Imperial College London
21
Kinematics (Gen) Jets
I dont know what this is
ORCA CMSSW
Y-axis N jets / event
21
Imperial College London
22
Kinematics (Rec) Jets
ORCA CMSSW
Y-axis N jets / event
22
Imperial College London
23
ResolutionObject ID cuts pT cuts applied but
no VBF background rejection cuts listed on page 5
appliedFor pT resolution, use?R(rec,mc)lt0.01
for leptons ? R(rec,mc)lt0.3 for jets
24
Resolution Muons and Electrons
muons corrected in CMSSW
check again after applying all the ele ID cuts
ORCA CMSSW
Y-axis of events
24
Imperial College London
25
Resolution Taujets and Tagging jets
no taujet correction yet for CMSSW
using sum of tracks for ORCA, calo position for
CMSSW (for now)
ORCA CMSSW
Y-axis of events
need to be investigated
25
Imperial College London
26
Resolution MET
ORCA CMSSW
Y-axis of events
26
Imperial College London
27
Mass Object ID cuts pT cuts applied but no
VBF background rejection cuts listed on page 5
applied(lack of statistics for CMSSW events)
28
Reconstructed Mass (Gen)
ORCA CMSSW
Y-axis of events
bug fixed but still some discrepancy
28
Imperial College London
29
Reconstructed Mass (Rec)
using calo MET (uncorrected)
ORCA CMSSW
Y-axis of events
using calo MET type 1
Some of this due to taujet not being corrected
for CMSSW not sure about this
29
Imperial College London
30
Vis-MET Separation (Gen)
ORCA CMSSW
Y-axis of events
?R, ?eta and ?phi between the sum of visible tau
decay products and MET (3-D missing momentum)
are quite different in ORCA and CMSSW
30
Imperial College London
31
Efficiency Object ID cuts pT cuts applied
but no VBF background rejection cuts listed on
page 5 appliedMatching of objects
?R(rec,mc)lt0.01 for leptons ? R(rec,mc)lt0.3 for
jets
32
Efficiency Lepton Taujet
Efficiency vs pT
Efficiency vs eta
Muon Electron Taujet
Muon Electron Taujet
lepton eta lt 2.5 taujet eta lt 2.3
lepton pT gt 10 GeV taujet pT gt 20 GeV
filled marker CMSSW hollow marker ORCA
This is a tagging efficiency ie. perform
offline search for a high pT obj and see if it
matches with MC Will check again with more
statistics for CMSSW and with all ID cuts applied
32
Imperial College London
33
Efficiency Jets Forward Jets
Efficiency vs pT
Efficiency vs eta
star jet reconstruction efficiency triangle
forward jet tagging efficiency
Different default pT cuts
filled marker CMSSW hollow marker ORCA CMSSW
a lot better!
eta lt 4.5
pT gt 20 GeV
Reconstruction Efficiency take a MC jet and see
if there is a matching rec jet Tagging Efficiency
perform offline forward jet tagging and see if
it matches with MC forward jets (bug fixed
since the last Higgs meeting)
33
Imperial College London
34
Slides for Higgs Meeting
35
Reconstruction in CMSSW vs ORCA
electrons muons
jets

• Comparison of reconstructed objects
• (qqH, H ? ?? ? lep jet with MH125GeV/c2,
• after the same offline kinematical cuts)
• ORCA (v873)
• CMSSW (v134)
• Performance better in CMSSW for most cases
• (full report at coming meetings)

missing ET (type1 muon)
35
Imperial College London
36
Efficiency Lepton and ?-jet
Efficiency vs pT
Efficiency vs ?
Muon Electron Taujet
Muon Electron Taujet
filled marker CMSSW hollow marker ORCA
lepton ? lt 2.5 ?-jet ? lt 2.3
lepton pT gt 10 GeV ?-jet pT gt 20 GeV
Object reconstruction tagging efficiency
perform offline search for a high pT obj and see
if it matches with MC (R lt 0.01 for leptons, R lt
0.3 for jets)
36
Imperial College London
37
Efficiency Jets Forward Jets
Efficiency vs pT
Efficiency vs eta
star jet reconstruction efficiency triangle
forward jet tagging efficiency
filled marker CMSSW hollow marker ORCA
? lt 4.5
pT gt 20 GeV
Reconstruction Efficiency take a MC jet and see
if there is a matching rec jet Tagging Efficiency
perform offline forward jet tagging and see if
it matches with MC forward jets
37
Imperial College London
38
Outlook on Reconstruction

• Concentrate on understanding and
• refining the reconstructed objects
• ? appreciate all the (new) tools in CMSSW
• Some topics
• ?-jets
• different ?-tagging methods
• (mass, IP, combined)
• PF
• Forward jets and Central Jet Veto
• JetVertexAssociation
• PF jets
• Track jets
• and more
• Development of analysis codes, HiggsAnalysis/VBFHi
  ggsTo2TauLJet/
• ? goal is to make a more standardised package

Reconstructed M?? (125GeV/c2) current status
ORCA CMSSW
basic cuts applied
38
Imperial College London