B Flavor Tagging Using Opposite Side Jet Charge 
Text for the blessed web page - CDF note 7131

The CDF Collaboration

August 13, 2004


We describe a Jet Charge algorithm for flavor tagging of B mesons. The algorithm determines the flavor of the signal B candidate by calculating the effective charge of the opposite side b jet. This version is our current best incorporating Silicon Stand Alone Tracking and 3 groups of different quality tags, including jets with high displaced track content. We measure the tagging power of the algorithm to be tex2html_wrap_inline179% on the inclusive e-SVT data sample of 290 pbtex2html_wrap_inline183.


This study measures performance of the Jet Charge Tagger (JQT) on a sample of semileptonic B decays. We use 290tex2html_wrap_inline185 of data collected with e-SVT and tex2html_wrap_inline189-SVT triggers. By applying subtraction in signed impact parameter of the displaced track to suppress light quark background and cutting on the invariant mass of the lepton and the displaced track to remove tex2html_wrap_inline191 contamination we achieve an equivalent of a 100% pure sample of semileptonic B decays.

The current version of the algorithm uses cone clustering algorithm to find track based jets. The cone size in tex2html_wrap_inline193 is 1.5 tuned for the best performance on data. As the first step of the algorithm all jets in each event are found on the opposite side from the trigger B meson. Next, the tagging jet is selected out of all available jets. There are three mutually exclusive groups of events.

  1. first, we look for a jet that contains a secondary vertex. These are the best, highest dilution jets, further referred as ``SecVtx jets''
  2. in the events with no SecVtx jet found we look for a jet that with the highest probability to contain displaced tracks. The probability for a jet to have high displaced track content is required to satisfy a certain cut. The events with such jets found fall into the second group called ``tex2html_wrap_inline195 jtes''
  3. finally, if an event has neither SecVtx nor tex2html_wrap_inline195 jets while does have *some* jets, the highest tex2html_wrap_inline199 jet is selected for tagging. These is the ``high tex2html_wrap_inline199 group.
After the tagging jet is selected, the jet charge tex2html_wrap_inline203 is calculated adding all tracks with weights proportional to thetex2html_wrap_inline199 of the tracks and the probability to be displaced. The sign of the jet charge determines the tagger decision.

The JQT is applied to the semileptonic B candidates. We measure the tagging power using the sign of the trigger lepton as an indicator for the trigger side B production flavor. The effects of mixing and sequential decays on the trigger side on the dilution are corrected for by a scale factor known from MC.

Blessed Numbers

This section summarizes the numbers which have been blessed for this analysis.

Tuned JQT performance on the e-SVT sample

The value for tex2html_wrap_inline211 is corrected for the trigger side mixing and sequential decays. It is obtained by adding subsamples of events with different quality tags. We have separated events into subsamples depending on whether the tagging jet contains a SevVtx, tex2html_wrap_inline195 or high tex2html_wrap_inline199 jet and into 11 bins of the jet charge tex2html_wrap_inline203 value.


Note: the error blessed is statistical only, however systematics have been estimated as well. Although we do not want to quote the values for systematics we have verified that the total systematic uncertainty is fairly close to the statistical one.

The full table below contains a more full picture for efficiencies and dilutions. The dilutions in the table correspond to the ``effective'' true dilution calculated as tex2html_wrap_inline219 where tex2html_wrap_inline211 is the combined value with all subsamples added.


The e-SVT sample used in this analysis contains tex2html_wrap_inline253484000 signal candidates after all cuts and background subtraction.

Blessed Figures

The following figures have been blessed for the mass difference measurement. More text can be found in CDF note 7131.

Figure 1: Jet charge distribution for the SecVtx, tex2html_wrap_inline195 and high tex2html_wrap_inline199 groups of jets for the e-SVT sample. Notice a better separation for jets with a secondary vertex.

Figure 2: Dilution dependence on the jet charge for the SecVtx, tex2html_wrap_inline195 and high tex2html_wrap_inline199 groups of jets for the e-SVT sample. Notice a generally higher dilution for jets with a secondary vertex.

About this document ...

B Flavor Tagging Using Opposite Side Jet Charge
Text for the blessed web page - CDF note 7131

This document was generated using the LaTeX2HTML translator Version 96.1-h (September 30, 1996) Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.

The command line arguments were:
latex2html blessed-jqt-lsvt.tex.

The translation was initiated by Ilya Kravchenko on Fri Aug 13 19:07:37 CDT 2004

Ilya Kravchenko

Fri Aug 13 19:07:37 CDT 2004