This preview shows page 1. Sign up to view the full content.
Unformatted text preview: misidentiﬁed as b jets while for the double-tag channel, a less
stringent requirement sufﬁces thereby giving increased per jet
tagging efﬁciency. The D0 single-tag analysis uses a b-tag
operating point which gives a typical efﬁciency of 48% with
misidentiﬁcation rate of 0.5% and an operating point for the
double-tag analysis which gives an efﬁciency of 59% and a
misidentiﬁcation rate of 1.7%. The CDF event selection
divides events into three exclusive categories based on which
algorithms identify a jet as b tagged. The ﬁrst category
contains events which have two jets tagged by the SECVTX
algorithm. The second category contains events not selected
into the ﬁrst category which have one jet identiﬁed by the
SECVTX algorithm and one by the jet probability algorithm.
The third category contains events which are not selected into
either of the other two categories and have one jet identiﬁed by
both the SECVTX and NN algorithms.
The sensitivity of the analyses is further improved by using
kinematic properties of the events to distinguish signal and
background events. The most important variable is the mass
of the dijet system corresponding to the Higgs decay. For
single-tagged events, the mass is computed using the tagged
jet and the highest pT jet remaining, and for the doubletagged events the mass is computed using the two tagged jets.
In addition to the mass, other variables can also distinguish
signal and background on a statistical basis. To make best use
of these, both experiments use a neural network to enhance
the separation of signal and background in the ﬁnal step of the
analysis. The CDF neural network has six input variables and
the D0 network has seven input variables.4 The variables used
by each experiment are listed in Table XXX for comparison.
Because of limited statistics in the W þ 3 jet sample, D0 uses
the NN only for the W þ 2 jet case and uses the dijet mass for
the W þ 3 jet case. Backgrounds to the WH search include
W þ jets production [including jets arising from heavy-ﬂavor
(hf) production], tt, single-top, diboson production (WW ,
WZ, and ZZ) and a small contribution from multijet events
in which either the lepton is actually a jet misidentiﬁed as a
lepton or the lepton arises from heavy-ﬂavor decay and the jet
energy is not reconstructed so the lepton appears to be
isolated. Both experiments estimate the tt, single-top, and
diboson yields using the product of the theoretical (N)NLO
cross section for each process, the luminosity and acceptance
times efﬁciency for each process. Corrections are applied 3
The event selection in recent preliminary results from CDF
include additional events in this channel which are selected by ET
Recent preliminary results also used the matrix element method
adapted from the top mass measurement and decision trees for the
multivariate technique instead of neural networks. mH (GeV=c2 )
180 ðgg ! H Þ (VBF) BðH ! WW Þ 54...
View Full Document
This document was uploaded on 09/28/2013.
- Fall '13