76
3 Jet Substructure at the LHC
Fig. 3.6 The CMS q/g
tagging performance in
simulation for three
individual observables and
their combined likelihood for
central jets with |η| < 1.3
and 80 < p T < 100 GeV.
Taken from [436]
gluon-jet rejection
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
quark-jet efficiency
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
qg-likelihood training variables:
ptD only
axis2 only
multiplicity only
All variables
< 100 GeV
T
| < 1.3, 80 < p
η
0 < |
CMS
Simulation Preliminary
(13 TeV)
dijets, Pythia8
uncertainties on the q/g tagging efficiency, samples of different but known quark
and gluon fractions are used to calibrate simulated events by correcting the corresponding distributions. ATLAS and CMS use Z /γ +jets and dijet samples, which
are enriched in quark and gluon jets, respectively. A template-based fit to the two
q/g likelihood distributions is used to determine weights for quark and gluon jets,
correcting the simulation for effects from mismodelling [436, 466]. Alternatively,
the rapidity dependence of the q/g fraction in dijet events can be used to extract the
track multiplicity separately for quarks and gluons [479]. The template fit results in
typical uncertainties of 2–5% for a wide range in p T , estimated on the same samples
as used for calibration. However, residual differences between data and simulation
are observed once the resulting calibrated tagger is applied to another final state,
which leads to additional uncertainties [478].
3.5.2 Vector Boson Tagging 7
The hadronic two-prong decays of weak vector bosons V have a distinct radiation
pattern compared to individual high- p T quarks or gluons. In particular, boosted
bosons tend to have two distinct subjets with relatively equal momentum sharing
(see Sect. 2.2.7). In contrast, most generic quark and gluon jets will have one prong
and if they have two, the second one tends to be soft. Furthermore, the mass of quark
and gluon jets scales with their p T and is lower than the electroweak boson masses
7 The text in this subsection has been taken from [26] and has been written by the author. It has been
adjusted to fit this book.
3 Jet Substructure at the LHC
Fig. 3.6 The CMS q/g
tagging performance in
simulation for three
individual observables and
their combined likelihood for
central jets with |η| < 1.3
and 80 < p T < 100 GeV.
Taken from [436]
gluon-jet rejection
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
quark-jet efficiency
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
qg-likelihood training variables:
ptD only
axis2 only
multiplicity only
All variables
< 100 GeV
T
| < 1.3, 80 < p
η
0 < |
CMS
Simulation Preliminary
(13 TeV)
dijets, Pythia8
uncertainties on the q/g tagging efficiency, samples of different but known quark
and gluon fractions are used to calibrate simulated events by correcting the corresponding distributions. ATLAS and CMS use Z /γ +jets and dijet samples, which
are enriched in quark and gluon jets, respectively. A template-based fit to the two
q/g likelihood distributions is used to determine weights for quark and gluon jets,
correcting the simulation for effects from mismodelling [436, 466]. Alternatively,
the rapidity dependence of the q/g fraction in dijet events can be used to extract the
track multiplicity separately for quarks and gluons [479]. The template fit results in
typical uncertainties of 2–5% for a wide range in p T , estimated on the same samples
as used for calibration. However, residual differences between data and simulation
are observed once the resulting calibrated tagger is applied to another final state,
which leads to additional uncertainties [478].
3.5.2 Vector Boson Tagging 7
The hadronic two-prong decays of weak vector bosons V have a distinct radiation
pattern compared to individual high- p T quarks or gluons. In particular, boosted
bosons tend to have two distinct subjets with relatively equal momentum sharing
(see Sect. 2.2.7). In contrast, most generic quark and gluon jets will have one prong
and if they have two, the second one tends to be soft. Furthermore, the mass of quark
and gluon jets scales with their p T and is lower than the electroweak boson masses
7 The text in this subsection has been taken from [26] and has been written by the author. It has been
adjusted to fit this book.