Results in ttHbb Channel Christian Shaw University of Glasgow

1
tth-gtbb neutralHiggs search channel
Christian Shaw University of Glasgow
2
Current Work - Artemis

• Reconstruction Comparisons within Athena with
  same code

FAST
FULL
ARTEMIS OBJECTS
ANALYSIS
Output Plots/Ntuples
3
Current Work - Artemis
2. Encourages re-usable Code and Object
Orientation
Analysis Code
Private
Artemis
Methods
Functors
Output Plots/Ntuples
4
Current Work - Artemis
MuID Propagates track parameters from MOORE back
to beam axis. Then parameterises scattering as
planes in calorimeters and calculates energy
loss. May also try ID (SCTTRT) track matching
with a c2 cut-off on difference.
MOORE Looks for regions of interest in f and RZ
projection. Performs Pattern recognition in MDT.
Track parameters (a0, z0, f, 1/pT) are expressed
at the first measured point

• Looking at leptons/ET only
• Atlfast e, m, missing ET
• Egamma e
• MOORE, MuID muons
• Reco Missing ET
• Raw data no, pT, h
• Leptons passing pT gt 20 GeV, h lt 2.5
• Reconstructed W ln (By solving for z component
  of the neutrino, with W mass constraint and set
  missing pT to n Px, Py)

5
Muon Results

MOORE and MuID Stand Alone find identical
numbers, as expected. Number of reconstruction
muons is much higher than with fast simulation.
Note that events with 1 only are roughly equal.
6
Muon Results

Most obvious point is the much larger number of
low pT objects in reconstruction. Note also the
MuID Standalone pT mean is slighter higher than
the Moore one
7
Electron Results

8
Lepton Results

• Saw all full reconstruction found more leptons,
  but of lower pT i.e. fake. What is difference
  after kinematic cuts?
  
• Not a lot of difference given statistics. About
  the same efficiency difference from egamma and
  muon reconstruction, compared to fast simulation.
  MuID appears to perform better (At least higher
  energies are found), presumably due to the
  calorimeteric calibrations.

9
WLEP Results

• Atlfast efficiency 50
• Egamma MOORE RecoET 47
• Egamma MuID StAl Reco ET 51
• Some difference, though resultant efficiencies
  are comparable. Is worth noting that the mean pT
  is somewhat higher for reconstruction, 125(120)
  GeV for Moore (MuID StAl) compared to 106 GeV for
  atlfast. With no jet information though it is
  unknown how this affects top reconstruction
  perhaps truth would help?

As relative to the single lepton efficiencies,
this translates to 76(AF), 82.5(MOO), 83.6(MuID
StAl)
10
Artemis Development

• Currently no MC truth information is available in
  Artemis.
• Adaptors need to be written to convert HEPMC
  objects to IParticle objects, which can be used
  for analysis.
• Probably a class will be defined for each
  particle, so user need not supply PdgID code.
  Question though as to when e.g. a lepton should
  be taken (After hard process, after radiation, at
  detector level).
• Also which comparisons would be valid at all
  (Quarks for instance are massless are should not
  really be compared with jets). However, W bosons
  e.g. should be OK.
• Code work is in progress, but these decisions are
  needed.

11
Future

• Would like to compare reconstructed quantities
  with truth, but what is sensible? (Distributions
  only, ratio of kinematics etc)
• B tagging is nearly ready in Athena, so if MC
  data works will be good experience for writing
  some jet flavour class/helper fn.
• Plan to compare the 3 hadronic decays in my
  channel when possible (h?bb ,W?jj, t ? bW),
  between AF and Full Sim
• Still interested in further investigating
  difference between alternate clustering
  mechanisms i.e. cone and kT, with a view to
  optimising for the tth channel

12
Webpage
Results online at http//ppewww.ph.gla.ac.uk/csh
aw/WEB/results.html