SLIDE 1
09/07/10 Matti Leinonen 1
09/07/10 Matti Leinonen 2
Overview of presentation
Track reconstruction
Measuring efficiency Changes made Efficiency plots
Track classification
Measuring efficiency Cuts used Efficiency table
T2 Track Reconstruction and Classification 09/07/10 Matti Leinonen - - PowerPoint PPT Presentation
T2 Track Reconstruction and Classification 09/07/10 Matti Leinonen - - PowerPoint PPT Presentation
T2 Track Reconstruction and Classification 09/07/10 Matti Leinonen 1 Overview of presentation Track reconstruction Measuring efficiency Changes made Efficiency plots Track classification Measuring efficiency Cuts used
SLIDE 2
SLIDE 3
09/07/10 Matti Leinonen 3
Information provided by Geant4
Geant4 provides information that can be used
to classify reconstructed tracks into primary and secondary tracks.
Primary tracks originate from the vertex. Secondary tracks originate from somewhere else.
Allows comparison of the number of simulated
tracks against the number of reconstructed tracks.
It is possible to study the effect of different cuts
to the classification of reconstructed tracks.
SLIDE 4
09/07/10 Matti Leinonen 4
Data used to generate plots
For the plots 10000 simulated 7 TeV MB pp
events were used.
Digitizer added Gaussian noise to the data. No dead vfats have been taken into account.
SLIDE 5
09/07/10 Matti Leinonen 5
Changes made to track reconstruction
Hit error
Moved from constant hit error to hit error based on
the number of clusters used in forming the hit.
Methods to calculate hit errors
Simulated information Leave-one-out method
Outlier point rejection
Previously no rejection for clearly erroneous hits in
track fitting.
Based on Chi-square and leave-one-out
calculations.
SLIDE 6
09/07/10 Matti Leinonen 6
Methods used for measuring the efficiency of track reconstruction
Chi-square distribution of reconstructed primary
and secondary tracks.
For all tracks there should be clear peak of tracks
near zero and constant number of tracks otherwise.
For reconstructed primary tracks the distribution
should be uniformly distributed.
The number of reconstructed primary tracks as
the function of simulated primary tracks.
The average number should be one with as small
deviation as possible.
SLIDE 7
09/07/10 Matti Leinonen 7
Reconstructed track Chi-square distribution
Before changes
SLIDE 8
09/07/10 Matti Leinonen 8
Reconstructed track Chi-square distribution
After changes
SLIDE 9
09/07/10 Matti Leinonen 9
Efficiency of track reconstruction
Before changes
SLIDE 10
09/07/10 Matti Leinonen 10
Efficiency of track reconstruction
After changes
SLIDE 11
09/07/10 Matti Leinonen 11
Cuts used in track classification
Chi-square cut
Track Chi-square value less than 1 %.
Old eta cut
Absolute value of track eta between 4.9 and 7.0.
Old Z cut
Absolute value of track Z less than 4000 mm.
New eta + Z cut
Based on the number of hits in the track and the eta
value of the track.
SLIDE 12
09/07/10 Matti Leinonen 12
Methods used in calculating the efficiency of track classification
Percentage of reconstructed primary and
secondary tracks that pass all track classification cuts.
Function , where p is the number of
reconstructed primary tracks that pass all classification cuts and s is the number of reconstructed secondary tracks that pass all classification cuts.
E p , s= p
s
SLIDE 13
09/07/10 Matti Leinonen 13
Effect of different cut combinations on reconstructed track classification
Chi-square Old Eta Old Z New eta + Z % of primary tracks pass cut % of secondary tracks pass cut E(p,s) 100.0 100.0 416 X 96.9 85.2 437 X 98.3 84.4 445 X 93.4 54.9 524 X 85.2 41.2 549 X X 93.4 54.9 524 X X 95.8 73.4 465 X X 91.8 52.0 530 X X X 91.8 52.0 530 X X 84.1 40.0 553
SLIDE 14
09/07/10 Matti Leinonen 14