Work notes on MET significance on data

MET significance - short definition.

The missing ET significance can be accessed by the MET::significance() function from CMSSW_3_4_X onwards for CaloMET and PFMET. It is defined as

$ S =  L(missing E_T)  / L(missing E_T =0) $

where the probabilities are calculated taking into account the uncertainties of all objects used in the calculation. In short, the MET significance is effectively the $\chi^2$ of the probability that the MET is consistent with a background fluctuation. (code can be found in CaloMET/METAlgorithms).


Here I document the cuts and preselection needed for data analysis.

  • cuts:
    • The EB quality cut is on the four closest (adjecent) crystals, and is called R4 = E(center)/E(5). It is not currently defined in any easy way for caloTowers. For now I used a bad event list ~apresyan/public/for_freya/runlist_ecalselection.txt
    • The HF has a cut on the photomultiplyer tubes being hit by particles and firing. This would result in no energy at all in one of the two PMTs that are used to read out the hcal cell. Some definitions: * the energy read out by the short fibre(S) is defined as 0.5*had energy, while the energy read out by the long fibre (L) is defined as the total EM energy - S.
        • So
           HF_S = 0.5* tower.hadEt();
           HF_L = tower.emEt()+HF_S;
           HFalpha = (HF_L-HF_S)/(HF_L+HF_S);
          reject tower if  -0.8<HFalpha<0.99
    • The primary vertex needs to be well-defined. The following cuts take care of this:
      •  !isFake()
  • Track quality for rejection of 'beam splash' (aka monster) events:
    •        count # tracks with quality "highPurity"
             reject events with #tracks>=10 AND fraction highPurity tracks<0.25

Analysis Flow

My analysis works in two steps:
  • re-run reconstruction and save reco output.
    • currently saved: outputCommands = cms.untracked.vstring("drop *","keep *_*_*_METSIGNIF","keep *_gtDigis_*_*","keep *_particleFlow_*_*","keep *_offlinePrimaryVertices_*_*","keep *_generalTracks_*_*")
  • Run a fwlite analysis on top of this. Macros can be found in the cvs UserCode repository.

-- FreyaBlekman - 26-Jan-2010

Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r5 - 2010-02-12 - FreyaBlekman
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Sandbox All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback