ACD Recon update, effects on background rejection

Using new ACD merit variables in background rejection

 Current background rejection

Prefilters:

ClassificationTree:

 Using New Variables

 AcdTkr1VetoSigmaHit and VetoHit1

AcdTkr1VetoSigmaHit is the probability of the "best" hit vetoing the track, expressed in "sigma".  This is the sum in quadrature of contributions from the miss distance and the signal being less that 1 MIP.  For MIP signals along the track projection this variable is 0.

 
Same plot, but with the X-axis remapped with log10( AcdTkr1VetoSigmaHit + 0.01)

For completeness here is a plot of  AcdTkr1VetoSigmaHit v. AcdTkr1ActiveDist

 
VetoHit1 is the same variable, including the tracker SSDVeto. 
VetoHit1 = sqrt(Acd2Tkr1VetoSigmaHit*Acd2Tkr1VetoSigmaHit+1.5*Tkr1SSDVeto*Tkr1SSDVeto)
Here is a plot of VetoHit1 

Same plot, but the X-axis remapped with log10(VetoHit1+0.01)

 
Note the small peak in the signal distribution around 7-9 is from low energy deposits in the tile pointed to by the track. 
 

  AcdTkr1VetoSigmaGap and VetoGap1

 AcdTkr1VetoSigmaGap is the probability that the track went in a gap in the ACD, expressed in "sigma".  This is the sum in quadrature of contributions geometrical  calculations of the gap and the chance of seeing no signal from the gap.   For high energy tracks that pass into the corner gaps of the ACD it is 0.  For tracks going into ribbons it is 1.5.

 Here is a plot of the low end of AcdTkr1VetoSigmaGap for all events:

 For comparison, here is a plot of AcdTkr1VetoSigmaGap v. AcdTkr1RibbonDist:

In analogy with VetoHit1 above we can define VetoGap1 to take into account the Trk1SSDVeto:
   VetoGap1 = sqrt(Acd2Tkr1VetoSigmaGap*Acd2Tkr1VetoSigmaGap+1.5*Tkr1SSDVeto*Tkr1SSDVeto)
 
Here is a plot of the low end of VetoGap1 for all events. 

If we look at only those events that pass weren't vetoed because of a track tile association or b/c of excess energy in the ACD we see this:
 

For Comparison, here is a plot of VetoGap1 v. AcdTkr1RibbonDist for the same set of events.

AcdTkrVetoSigmaHit and VetoHit

We can also select the best association by looping over all tracks.
Here is a plot of that variable, AcdTkrVetoSigmaHit: 

 
In analogy to VetoHit1 we can make VetoHit from  AcdTkrVetoSigmaHit. 
Here is a plot of that:
 

 
Then if we only look at events that weren't rejected because of VetoHit1 or excess energy:

 
So, we could consider cutting very loose on this and rejecting some event with a high probablity match.
 

Reproducing Existing Performance with simple cuts 

Some other useful new Variables, (for Classification Tree)

 Another type of variable we consider is the scaled energy in a cone around the track.  We have three variables that reflect 15, 30 and 45 degreee cones.

With these variables (and some familiar ones) we can make a Classification Tree.

This plot show the performance of the CT (in red) relative to the pass 6 cuts (3 black circles).  The prefilter performances are also shown.
For comparison the CT performance without any prefilters is also show (in blue). 
 
Just to give some numbers, here are numbers for 3 difference values of cutting on the CT output. 
|| Class || Gamma Effic || Bkg Effic
|| Expression
||

 Basically, these numbers correspond to either 20% reduction in residual background at given signal efficiency or 4% increse in signal efficiency at give background.