Mistag calibration with multijet data

Complete: 4

Method Description

The present study relies on tracks with negative impact parameters in multijet QCD data. For each jet the b-tag discriminator(s) can be computed with these tracks only, thus giving a jet sample with reduced b and c quark contamination (around a few percents). Of course for the same jet, the standard b-tag discriminator can also be used (relying on all negative+positive tracks for the time being). The MC can be used to estimate the ratio R_light = efficiency of standard udsg tag / efficiency of udsgcb negative tag, where the efficiencies are the ratio: number of tagged jets / number of taggable jets (here, for the time being, a taggable jet is simply a jet with at least 2 associated tracks). Thus in multijet real data, the mistags can be estimated as the product: effi_light = efficiency of negative tag * R_light where effi_light can be parametrised versus pt, eta, number of pile-ups, etc... Without applying the tagger in the MC, effi_light can be directly used as a weight in any udsg quark jet MC.

Another way (which I prefer) is to compute a scale factor SF_light = efficiency of negative tags in multijet real data / efficiency of udsgcb negative tag in QCD MC. After applying the tagger in MC, SF_light can be applied as a weight on each udsg jet, here also depending on jet pt, eta, PU, ...

At present this method has been tested with the track counting method on CMSSW 1.2.0 QCD samples within 30<pt<230 GeV. It will be also tested with the probabilistic method (as negative track informations should be available with 1.3.0).

Systematics have been evaluated by varying the b, c and g fractions in the QCD MC.

List of tasks and todo

1 Porting

Currently working: Daniel Bloch, Jérémy Andrea
Description: Move the code to compute R_light and systematics in an Analyser
Status: to be done (with 1.3.0 ?)

2 Adapt the secondary vertex tagger and combined tagger for negative tracks !

3 Study the effect of tracks from K0s and Lambda decays, and from photon conversions

4 Use photon+jet events rather than di-jet events (less b, c contamination ?)

5 Use of light jets (from W decays) in ttbar events

6 Sensitivity to fragmentation parameters in MC generators


Main.blochd - 15 Feb 2007 - page author

Responsible: Main.blochd

Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r3 - 2007-09-25 - JennyWilliams



 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback