Update on track reconstruction in the Near Detector

1
Update on track reconstruction in the Near
Detector

• N. Saoulidou, Fermilab 1-29-04

2
Outline

• Goal and strategy.
• Comparison between my new version of the code
• and the existing one using
• My simple code
• Truth Helper
• Summary On going work

3
Goal Strategy

• Goal
• Reconstruct the muon track from the vertex to
  the spectrometer region.
• Strategy
• a) Assume that event slicing is already done
  (process only single events were things are much
  more clear).
• b) Correctly include all 4 different strip
  solutions with equal weights for the spectrometer
  region (CandStripSR) and then proceed with the
  tracking tuning various parameters to account for
  the specific characteristics if the Near
  Detector.
• After very useful discussions with Jim I
  managed to create a correct list of Candidate
  Strips for the spectrometer region (as well),
  that contain list of digits (my previous code
  was not handling this issue in a proper way).

4
Examples of reconstructed muon tracks cont.
5
Quantitative results of the track reconstruction
(using my simple code for calculating the track
quality and track reconstruction efficiency ).

• Percentage() of the reconstructed muon track
  strips that belong to the true muon track (for
  5000 NEW Near Det. MC events and DetSim ), after
  considering
• a) numu CC events with the neutrino vertex up to
  20 plates upstream the spectrometer region
  (just to make sure that the event is far enough
  from the spectrometer region) and
• b) the muon track to have more than 5 planes
• The mean value of this distribution is 85 for
  my code and 82 for the current code.
• The percentage of numu CC events satisfying the
  previous criteria were no track was reconstructed
  is 20 with my code and 30 with the current
  code.(80 of the mu CC events are reconstructed
  with my code and 70 with the current code)

Current code (Jims code)
My code
6
Quantitative results of the track reconstruction
(using my simple code for calculating the track
quality and track reconstruction efficiency ).

• The difference of the reconstructed momentum
  (using the SR fitter) and the actual muon
  momentum shows for both cases a shift which is
  slightly more pronounce using my code
  (statistics?).

SAME Fitter
Current code
My code
7
Quantitative results of the track reconstruction
(using Truth Helper).

• The track purity is the same between the two
  versions.

Current code
My code
8
Quantitative results of the track reconstruction
(using Truth Helper) contd.

• The track completeness is much better with the
  current code BUT.
• This is contradicting (?) the results obtained
  with my simple code for calculating the track
  quality
• What I calculate is
• I would assume that this would correspond to the
  track completeness as calculated by the Truth
  Helper but apparently it is not ( need to
  investigate that)

Current code
My code
9
Quantitative results of the track reconstruction
(using Truth Helper) contd.

• The track completeness (slc) is again better with
  the current code and in both cases it improves
  very little in comparison with the previous track
  completeness.

My code
Current code
10
Summary On going work

• The track reconstruction efficiency (number of
  reconstructed tracks) of my version of the code
  is higher, as shown by both the Truth Helper and
  my simple code for calculating track
  reconstruction characteristics.
• The track purity of both version of track
  reconstruction appears to be the same using both
  Truth Helper and my simple code (for calculating
  track reconstruction characteristics.)
• However the track completeness (which is
  proportional to what I am calculating
  
  )
• A) Is very different between the Truth Helper
  and my simple code and this needs to be
  investigate
  
• B) Appears to be much better with the current
  version of the track reconstruction
• In order to come to a conclusion (related with
  which track reconstruction code is more
  efficient) I need to understand better the Truth
  Helper variable track completeness .