track-based electron shower algorithm for improved electron PID
Adds an algorithm that calculates track-based electron shower variables to improve electron PID. Basically, it estimates the energy distribution in surrounding calo cells and uses this to calculate discriminating variables.
It can replace EcalEnergyForTrack and calculates more and better performing variables for marginal extra speed costs.
Energy distributions are estimated using MC simulation with ideal EM shower shapes (see RevModPhys.75.1243), translated into cell/track parameters for LHCb. Essentially comes down to a parametrization/look-up table for cell energies, assuming it is an electron.
Expectation values can be used to select cells for E/p and calculate a summed per-cell-E/p DLL (improvement from chi2 like variable calculated in earlier versions of this algo).
Details in link-to-pdf (see also link-to-older-pdf). Update on going from chi2-like variable to summed DLL of E/p per-cell: link-to-pdf
Needs TH3 support in Gaudi and LHCb, see: Gaudi!1226 and LHCb!3112 (merged)/LHCb!3208 (merged)
First was parametrization was in CondDB: see SIMCOND!152. Now moved to ParamFiles: ParamFiles!35.
Moore implementation here: Moore!901 (merged)
Merge request reports
Activity
added Calo PID enhancement labels
added new feature label
- Resolved by Maarten Van Veghel
- Resolved by Maarten Van Veghel
mentioned in issue #195 (closed)
added 415 commits
-
d4823272...03d5a841 - 414 commits from branch
master - e5266303 - Merge remote-tracking branch 'origin/master' into mveghel-electronshowervars-dev
-
d4823272...03d5a841 - 414 commits from branch
added 1 commit
- 76d1bcb6 - MAJOR OVERHAUL: energy expectation param based on MC study
added 1 commit
- b7ecf49a - energy to eoverp; some speed improvements; default no TH3 interpolation
added 1 commit
- 0979e3ed - many tweaks; pid/sel optimization; default hybrid interpol+bin look-up
