GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » Tracking » eta_c reconstruction efficiency
eta_c reconstruction efficiency [message #12794] Sat, 19 November 2011 23:22 Go to next message
Gianluigi Boca is currently offline  Gianluigi Boca
Messages: 177
Registered: March 2004
first-grade participant
From: *ibz.tu-darmstadt.de
Hello,
I verified that the loss in efficiency in the Eta_c channel certainly depends HEAVILY on the selection criterion I used in the FINAL association (ie, after the track parameters are found by the STT + Mvd Pattern Recognition) of the Mvd hits to the tracks found by the pattern recognition.

The stricter criterion, used in the July11 version, gives the best efficiency.

I restored this stricter selection to the PndSttMvdTracking.cxx version presently in the trunk.

Therefore now Dima et al. can start reanalysing the data and see what happens.

From my side I'd like to continue to investigate :
1) whether this is the only cause of the drop in efficiency;
2) understanding if it is the Kalman that is very sensitive to the presence of one or two spurious hits or if it is the vertexing algorithm or some thing else.

Cheers ! Gianluigi
Re: eta_c reconstruction efficiency [message #12795 is a reply to message #12794] Mon, 21 November 2011 10:07 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Which are the selections cutting so strongly the efficiency? Because, as far as I remember, also Elisa and Marius analyses were suffering from this drop, and I suppose they have used different selection criteria. Also I was able to see the drop, with my macro completely different from the Dima's one, and without using cuts (apart from the mc id).
Re: eta_c reconstruction efficiency [message #12796 is a reply to message #12794] Mon, 21 November 2011 11:15 Go to previous messageGo to next message
Dima Melnychuk is currently offline  Dima Melnychuk
Messages: 213
Registered: April 2004
Location: National Centre for Nucle...
first-grade participant
From: *fuw.edu.pl
Hi Stefano,

In my understanding the problem was not with selection in analysis but with cut used internally in PndSttMvdTracking.cxx
You can see inside PndSttMvdTracking.cxx now:

const Double_t MvdCut=0.3;

insead of

const Double_t MvdCut=0.8;

So the data should be reproduced at least from the reconstruction level (and I suppose for all the channels).

Dima
Re: eta_c reconstruction efficiency [message #12797 is a reply to message #12794] Mon, 21 November 2011 11:30 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Sorry, I had understood that we were speaking about the analysis selection critera.
I will check the new trunk, then.
Re: eta_c reconstruction efficiency [message #12802 is a reply to message #12797] Mon, 21 November 2011 17:46 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
I have tried the new trunk and compared the eta peak with the old trunk and the "good" r12837 revision:

index.php?t=getfile&id=6728&private=0

The improvement in the new trunk is quite visible comapred top the old trunk! The peak seems a bit lower than july version, but we are quite close to the good value!
  • Attachment: newcomp.gif
    (Size: 9.59KB, Downloaded 405 times)
Re: eta_c reconstruction efficiency [message #12804 is a reply to message #12802] Tue, 22 November 2011 14:52 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
I have also run 1000 eta_c events, and compared the results of the new trunk with (red) and without (blue) cleanup.

index.php?t=getfile&id=6729&private=0

Maybe it is time to check higher statistics on the grid.
Re: eta_c reconstruction efficiency [message #12808 is a reply to message #12804] Wed, 23 November 2011 17:59 Go to previous message
Gianluigi Boca is currently offline  Gianluigi Boca
Messages: 177
Registered: March 2004
first-grade participant
From: *gsi.de
Stefano Spataro wrote on Tue, 22 November 2011 14:52

I have also run 1000 eta_c events, and compared the results of the new trunk with (red) and without (blue) cleanup.

index.php?t=getfile&id=6729&private=0

Maybe it is time to check higher statistics on the grid.



Hi,
yes, I know, there is still a small inefficiency (7%) in the TRUNK relative to the July11 version .

This again is due to some extra spurious hits, this time in the STT
system.
Again it is most likely due to some more relaxed criterion in the Pattern Recognition, in the association of the of the STT hits to a found track.

Therefore I am fixing this for the time being by restoring the older stricter criterion.

The next step will be to investigate where exactly these few extra spurious hits have this negative impact (Kalman filter? Vertexing routines ? Analysis selection criteria ?).

My feeling today is that when we have an operating Deterministic Annihiling Kalman filter (which is already in place, but not the default) these cases could be overcome.

Gianluigi
Previous Topic: track quality criterion
Next Topic: Eta_c efficiency
Goto Forum:
  


Current Time: Mon Nov 11 07:30:57 CET 2024

Total time taken to generate the page: 0.00706 seconds