Status of track reconstruction in the Near Detector

1
Status of track reconstruction in the Near
Detector

• N. Saoulidou, Fermilab 2-26-04

2
Outline

• Corrections of ND tracking
• Results (using Truth Helper)
• Summary On going work

3
Corrections on ND tracking

• Modified the way duplicate 3D tracks are
  examined
• Initially the percentage of identical hits was
  examined for all track strips (U and V together).
• There were some cases were false 3D solutions
  existed having the correct 2D Line in one view
  and the wrong 2D Line in the other view.
• The code is now modified to examine percentage of
  identical strips in each view and keep the
  solution with the largest number of planes.
• Corrected the plane number from which the Plex
  starts from gt 121 to gt121!
• Due to that error a few quite clear tracks were
  missed because the code was unable to reconstruct
  the correct 2D line in the U view.

4
Results of the track reconstruction after
corrections (Track Completeness for all tracks
using Truth Helper).
Before
After
Mean 0.75
Mean 0.70
Mean 0.86
Mean 0.88

• The track completeness improved
• Less events missed (zero completeness)
• The completeness of the ones that are
  reconstructed has increased

5
Results of the track reconstruction (Track
Completeness for muon tracks using Truth Helper).
Before
After
Mean 0.88
Mean 0.90
Muons

• The track completeness for muons has also
  increased.

6
Results of the track reconstruction after
corrections (Track Purity for all tracks using
Truth Helper).
Before
After
Mean 0.83
Mean 0.80
Mean 0.87
Mean 0.90

• The track purity aslo improved
• Less events missed (zero purity)
• The purity of the ones that are reconstructed has
  increased

7
Results of the track reconstruction (Track Purity
for muon tracks using Truth Helper).
Before
After
Mean 0.96
Mean 0.96
Muons

• The track purity for muons is the same (very
  small increase)

8
Things to be investigated

• How is it possible the purity to be non zero and
  the completeness to be exactly zero
• Why the Truth Helper variable completeslc (how
  the slicing affects the results) is not filled
  any more

9
Summary On going work

• Making some corrections of the ND tracking the
  completeness and the purity of the reconstructed
  tracks has increased quite significantly and I
  would like to add these changes in the offline
  code.
• There are a few more things related with ND track
  finding that need to be investigated.
• I will try to learn how to use the PerfTools (
  for the moment they are not available at fnalu)
  in order to investigate why the ND track fitting
  takes so much time