Search for the standard model Higgs boson produced in association with top quarks in multilepton final states

This is a condensed description with plots for the analysis HIG-13-020.

Abstract

A search for the standard model Higgs boson produced in association with a top quark pair is presented, using 19.6 fb-1 of 8 TeV pp collision data collected by the CMS experiment at the LHC. Final states with a Higgs boson that decays to either ZZ*, WW*, or ττ are required to have a top quark pair that decays to either lepton plus jets (tt → ℓvjjbb) or dileptons (tt → ℓvℓvbb), where ℓ represents an electron or a muon. The following signatures are selected: two isolated same-sign leptons (electrons or muons) plus b-tagged jets, three isolated leptons plus b-tagged jets, or four isolated leptons plus b-tagged jets. The expected 95% confidence level upper limit on the Higgs boson production cross section for a Higgs boson mass of 125.7 GeV/c2 is 2.4 times the standard model expectation, to be compared to an observed limit of 6.6. The signal strength μ, relative to the expectation for the standard model Higgs boson, is measured to be μ =3.7
+1.6
–1.4
.

Main plots

CMS Combination

This analysis is included in the CMS combination of search results for Higgs boson production in association with a top-quark pair. Documentation for the combination is available here.

Plot Caption
The best-fit values of the signal strength parameter μ = σ/σSM for each ttH channel at MH = 125.7 GeV . The signal strength in the four-lepton final state is not allowed to be below approximately 6 by the requirement that the expected signal-plus-background event yield must not be negative in any of the two bins of jet multiplicity.
The observed and expected 95% CL upper limits on the signal strength parameter μ = σ/σSM for all ttH channels combined, as a function of the mass of the Higgs boson.

Jet multiplicity and BDT discriminant distributions

Common caption In these plots events with positive and negative charge are merged. The signal yield is the amount predicted by the standard model (μ = 1). The background yields are from the combined fit to the final discriminant at fixed μ = 1. The bottom panel of each plot shows the ratio between the observed events and the expectation from simulation, with statistical and systematical uncertainties on the expectations after the fit.

Plot Caption
Jet multiplicity distribution for the e±e± final state.
The last bin of the distribution includes also events with higher jet multiplicity.
Jet multiplicity distribution for the e±μ± final state.
The last bin of the distribution includes also events with higher jet multiplicity.
Jet multiplicity distribution for the μ±μ± final state.
The last bin of the distribution includes also events with higher jet multiplicity.
BDT output distribution for the trilepton final state.
The last bin of the distribution includes also events with higher jet multiplicity.
BDT output distribution for the four-lepton final state.
The last bin of the distribution includes also events with higher jet multiplicity.
BDT output distribution for the e±e± final state.
BDT output distribution for the e±μ± final state.
BDT output distribution for the μ±μ± final state.
BDT output distribution for the trilepton final state.

Signal extraction results

Plot Caption
Best fit values of the signal strength parameter μ = σ/σSM and ±1σ uncertainties, for the five individual final states (solid markers with red error bars) and the full combination (vertical line and green band). The signal strength in the four-lepton final state is not allowed to be below approximately 6 by the requirement that the expected signal-plus-background event yield must not be negative in any of the two bins of jet multiplicity).
Best fit values of the signal strength parameter μ = σ/σSM and ±1σ uncertainties, for the three final states (solid markers with red error bars) and the full combination (vertical line and green band). The signal strength in the four-lepton final state is not allowed to be below approximately 6 by the requirement that the expected signal-plus-background event yield must not be negative in any of the two bins of jet multiplicity).
95% CL upper limit on the signal strength parameter μ = σ/σSM for the three final states and their combination: observed (solid markers), median expected under the background-only hypothesis (hollow markers), and intervals containing 68% and 95% of the expected outcomes under that hypothesis (green and yellow bands).
95% CL upper limit on the signal strength parameter μ = σ/σSM for the five individual final states and their combination: observed (solid markers), median expected under the background-only hypothesis (hollow markers), and intervals containing 68% and 95% of the expected outcomes under that hypothesis (green and yellow bands).
95% CL upper limit on the signal strength parameter μ = σ/σSM for the combination of all final states, as a function of the Higgs boson mass: observed (solid markers), median expected under the background-only hypothesis (hollow markers), and intervals containing 68% and 95% of the expected outcomes under that hypothesis (green and yellow bands).

Support slides on the cross-checks

Some slides describing briefly all the cross-checks done to investigate the result, in particular in the dimuon final state, are available in PDF PDF and PPTX (2007) PowerPoint formats.

Extra plots: Distributions of the kinematic variables used as input in the BDT (not in the PAS)

Dilepton final state

Common caption: In the following plots, unlike in the BDT output and jet multiplicity ones, the background yields are determined from control regions and simulations alone, without fitting the data also in the signal region. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).

Plot Caption
HT, the scalar sum of the transverse momenta of the leptons and jets in the event, for the e±e± (left), e±μ± (center) and μ±μ± final state.
HTmiss, the magnitude of the vector sum of the transverse momenta of the leptons and jets in the event, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Trailing lepton transverse momentum, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Trailing lepton pseudorapidity, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Angular separation between the trailing lepton and the closest hadronic jet, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Transverse mass of the leading lepton – ETmiss system, for the e±e± (left), e±μ± (center) and μ±μ± final state.

Trilepton final state

Common caption: In the following plots, unlike in the BDT output and jet multiplicity ones, the background yields are determined from control regions and simulations alone, without fitting the data also in the signal region. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).

Plot Caption
HT, the scalar sum of the transverse momenta of the leptons and jets in the event.
Fraction of HT from objects within |η|<1.2 with respect to the total HT.
pT of the jet with highest b-tagging discriminator.
Maximum |η| among the three leptons.
ΔR between the two closest opposite-sign leptons.
Invariant mass of the reconstructed hadronically-decaying top quark candidate. The bin at zero contains the events for which no hadronically-decaying top quark candidate is reconstructed (events with only two jets, or three jets of which two are b-tagged).

Extra plots: Cross checks (not in PAS)

Di-muon final state BDT output distributions with different selections

Plot Caption
BDT output distribution for the μ±μ± final state using a cut-based muon selection as in CMS PAS SUS-13-013.
BDT output distribution for the μ±μ± final state using a looser MVA muon selection, inclusive of the nominal muon selection (left plot) or exclusive to it, i.e. not selecting the events for which both muons pass the nominal selection (right plot).
This working point corresponds to an expected reducible background yield approximately twice the one expected for the nominal selection.
BDT output distribution for the μ±μ± final state using a looser MVA muon selection, inclusive of the nominal muon selection (left plot) or exclusive to it, i.e. not selecting the events for which both muons pass the nominal selection (right plot).
This working point corresponds to an expected reducible background yield approximately four times the one expected for the nominal selection.
BDT output distribution for the μ±μ± final state using a looser MVA muon selection, inclusive of the nominal muon selection (left plot) or exclusive to it, i.e. not selecting the events for which both muons pass the nominal selection (right plot).
This working point corresponds to an expected reducible background yield approximately six times the one expected for the nominal selection.
BDT output distribution for the μ±μ± final state in the sideband of events with only three hadronic jets (the signal selection requires at least four).

Distributions of kinematic and lepton identification variables in the dimuon final state

Common caption: These plots show the distribution of some selected kinematic and lepton-identification variables for the dimuon final state, after the full event selection.
The prediction for the reducible background yield is derived from control regions and a fit to the BDT output distribution at fixed μ=1, while the prediction for the shape is from simulations, as the data-driven method is not applicable to the distributions of the lepton identification variables. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).
The purpose of these plots is to illustrate that the kinematic of the observed excess of events is compatible with that of prompt leptons (ttV, ttH) and that no anomalies are seen in the distributions.

Plot Caption
Scalar sum of the transverse momenta of the two leptons.
Invariant mass of the two leptons.
Angular separation between the two leptons.
Sum of the charges of the two leptons.
HT, the scalar sum of the transverse momenta of all the leptons and jets.
ETmiss.
Transverse mass of the leading lepton – ETmiss system.
Number of b-jets satisfying the medium working point of the CSV tagger.
Transverse momentum of the trailing jet among the two with highest b-tagging discriminator value in the event.
Number of leptons in the event after the preselection, before the lepton requirement.
Lepton MVA discriminator value for the less signal-like among the two muons.
The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.
Relative isolation value for the less isolated among the two muons.
Largest absolute transverse impact parameter among the two muons.
Largest absolute longitudinal impact parameter among the two muons.
Largest 3D impact parameter significance among the two muons.

Distributions of kinematic and lepton identification variables in the dimuon final state, with relaxed lepton MVA requirements

Common caption: These plots show the distribution of some selected kinematic and lepton-identification variables for the dimuon final state, after the full event selection but using a looser working point for the lepton MVA, corresponding approximately to a 5-fold increase in the expected reducible background. The distributions are presented both for all events, and for the events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
The prediction for the reducible background yield is derived from control regions and a fit to the BDT output distribution at fixed μ=1, while the prediction for the shape is from simulations (as the data-driven method is not applicable to the distributions of the lepton identification variables). The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).
The purpose of these plots is to check the distributions of the events in a sample that contains a larger contribution from reducible background compared to the nominal selection, and that the BDT discriminator does properly select events that are more signal-like.

Plot Caption
Scalar sum of the transverse momenta of the two leptons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Invariant mass of the two leptons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Angular separation between the two leptons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Sum of the charges of the two leptons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
HT, the scalar sum of the transverse momenta of all the leptons and jets. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
ETmiss. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Transverse mass of the leading lepton – ETmiss system. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Number of b-jets satisfying the medium working point of the CSV tagger. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Transverse momentum of the trailing jet among the two with highest b-tagging discriminator value in the event. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Number of leptons in the event after the preselection, before the lepton requirement. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Lepton MVA discriminator value for the less signal-like among the two muons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.
Relative isolation value for the less isolated among the two muons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Largest absolute transverse impact parameter among the two muons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Largest absolute longitudinal impact parameter among the two muons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
Largest 3D impact parameter significance among the two muons. The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.

Lepton MVA output distribution in the dimuon final state

Common caption:
Plot Caption
Distribution of the lepton MVA discriminator for events satisfying the μ±μ± selection state except for the lepton MVA requirement, for the most background-like of the two muons. High values of the discriminator correspond to signal-like leptons, and the requirement of the nominal selection is MVA > 0.7 for both leptons.
The plot on the left is for all the events, the plot on the right only for events with a signal-like kinematic, selected by requiring the final BDT discriminator value to be larger than 0.2.
In this plot, the expected distributions for all signal and background processes are derived from simulation, but the yield for the reducible background is fitted from the data from this distribution.
The purpose of this plot is to show that, in the dimuon final state, except for the excess of events observed at high lepton MVA values the distribution of the lepton MVA discriminator is well modelled by the known reducible backgrounds.

Lepton identification variables in opposite-sign dimuon events from leptonic ttbar decays

Common caption: These plots show the distribution of some selected lepton-identification variables for the dimuon final state after the full event selection, except that the two muons are required to be opposite-sign in order to select pairs of prompt dimuons from fully leptonic ttbar decays. To reduce the contribution from Z→ll, events for which the dilepton mass is within 10 GeV from the nominal Z mass are vetoed, and a loose requirement on the ETmiss likelihood discriminator.
The predictions for the distributions are taken from simulations, with corrections derived from Z→ll events in data, but the overall normalization of the tt is scaled to match the observed yield in data. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties.
The purpose of these plots is to validate the modelling of identification variables from prompt leptons in events with similar multeplicities of jets and b-jets to the signal ones.

Plot Caption
Lepton MVA discriminator value for the less signal-like among the two muons.
The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.
Relative isolation value for the less isolated among the two muons.
Ratio of the lepton pT to the associated jet pT (input to the lepton MVA), for the less signal-like of the two muons.
b-tagging discriminator of the jet associated to the lepton (input to the lepton MVA), for the less signal-like of the two muons.
Relative isolation value for the less isolated among the two muons.
Largest absolute transverse impact parameter among the two muons.
Largest absolute longitudinal impact parameter among the two muons.
Largest 3D impact parameter significance among the two muons.

Muon impact parameter distributions in other final states

Common caption: These plots show the distribution of the impact parameter variables of muons for events in final states other than the dimuon signal region. The prediction for the reducible background yield is derived from control regions and a fit to the BDT output distribution at fixed μ=1, while the prediction for the shape is from simulations (as the data-driven method is not applicable to the distributions of the lepton identification variables). The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).
The purpose of these plots is to validate these variables, and exclude the presence of sizeable unaccounted backgrounds with non-primary muons.

Plot Caption
Absolute transverse impact parameter of the muon, for e±μ± events after the full selection. In this final state, approximately 40% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
Absolute longitudinal impact parameter among all the muons, for e±μ± events after the full selection. In this final state, approximately 40% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
3D impact parameter significance among all the muons, for e±μ± events after the full selection. In this final state, approximately 40% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
Largest absolute transverse impact parameter among all the muons, for trilepton events after the full selection. The variable is defined to be zero for events that have only electrons. In this final state, approximately 45% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
Largest absolute longitudinal impact parameter among all the muons, for trilepton events after the full selection. The variable is defined to be zero for events that have only electrons. In this final state, approximately 45% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
Largest 3D impact parameter significance among all the muons, for trilepton events after the full selection. The variable is defined to be zero for events that have only electrons. In this final state, approximately 45% of the reducible background is from events with a non-prompt muon, the remaining from events with a non-prompt electron.
Largest absolute transverse impact parameter among the two muons, for μ±μ± with exactly three hadronic jets, but otherwise satisfying the full dimuon selection.
Largest absolute longitudinal impact parameter among the two muons, for μ±μ± with exactly three hadronic jets, but otherwise satisfying the full dimuon selection.
Largest 3D impact parameter significance among the two muons, for μ±μ± with exactly three hadronic jets, but otherwise satisfying the full dimuon selection.

Statistical analysis results

Common caption: These plots show the result of the statistical analysis when doing the signal extraction using the multiplicity of hadronic jets instead of the BDT output, and when using a cut-based lepton selection instead of the lepton MVA. In these checks the four-lepton channel is left unmodified as the signal extraction there is done using the jet multiplicity even in the nominal analysis, and the tight cut-based selection would not be appropriate due to the low signal yield.
The purpose of these plots is to check quantitatively that the results derived with simpler, though less performant, approaches.

Plot Caption
Signal extraction results using as discriminating variable the multiplicity of hadronic jets instead of the BDT output in the dilepton and trilepton final states
95% CL upper limit on the signal strength parameter μ = σ/σSM for the five individual final states and their combination: observed (solid markers), median expected under the background-only hypothesis (hollow markers), and intervals containing 68% and 95% of the expected outcomes under that hypothesis (green and yellow bands).
Signal extraction results using as discriminating variable the multiplicity of hadronic jets instead of the BDT output in the dilepton and trilepton final states
Best fit values of the signal strength parameter μ = σ/σSM and ±1σ uncertainties, for the five individual final states (solid markers with red error bars) and the full combination (vertical line and green band). The signal strength in the four-lepton final state is not allowed to be below approximately 6 by the requirement that the expected signal-plus-background event yield must not be negative in any of the two bins of jet multiplicity).
Signal extraction using the cut-based lepton selection in the dilepton and trilepton final states.
Best fit values of the signal strength parameter μ = σ/σSM and ±1σ uncertainties, for the five individual final states (solid markers with red error bars) and the full combination (vertical line and green band). The signal strength in the four-lepton final state is not allowed to be below approximately 6 by the requirement that the expected signal-plus-background event yield must not be negative in any of the two bins of jet multiplicity).

Signal extraction with unconstrained normalizations of the main backgrounds yields

As a cross-check, the signal extraction procedure is repeated relaxing the constraints on the main backgrounds. The fit to the data is performed with five floating parameters: three signal strength parameters for the ttH, ttW, and ttZ processes, and two signal strength parameters for the reducible backgrounds from non-prompt muons and non-prompt electrons.
To better separate the ttH signal from the ttV backgrounds, additional control regions are included in the fit as separate categories: same-sign dilepton events with exactly three jets, mainly from ttW and reducible backgrounds, and trilepton events with one on-shell Z candidate, mainly from ttZ production.

The result from the fit is μ(ttH) = 2.8
+1.8
–1.6
, the difference with the nominal result being compatible with the loss in sensitivity at the 1σ level.
The fitted signal strengths for the backgrounds are also all compatible with the input predictions within 1σ or better: μ(ttW) = 1.4
+0.6
–0.5
, μ(ttZ) = 1.1
+0.4
–0.3
, μ(fake μ) = 0.7
+0.4
–0.3
, μ(fake e) = 0.9
+0.3
–0.3
. The most important bidimensional correlations are shown in the plots below.

Plot Caption
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttH) and μ(ttW), profiling the remaining parameters.
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttH) and μ(ttZ), profiling the remaining parameters.
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttH) and μ(fake μ), profiling the remaining parameters.
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttH) and μ(fake e), profiling the remaining parameters.
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttW) and μ(fake μ), profiling the remaining parameters.
Best fit value (cross), 68% CL (solid), and 95% CL (dashed) contours obtained from a two dimensional likelihood scans in μ(ttW) and μ(fake e), profiling the remaining parameters.

Extra plots: Control regions (not in PAS)

Dilepton final state: events with one lepton failing the lepton MVA

Common caption:These plots show the distribution of the main discriminating variables and the final BDT output for events where one lepton fails the lepton MVA requirement, that are then used to determine the reducible background predictions in the signal region.
In these plots, the expected distribution for the reducible background is taken from simulations, and the yield is fitted from the data. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties after the fit (light blue).

Plot Caption
BDT discriminator output, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Jet multiplicity, for the e±e± (left), e±μ± (center) and μ±μ± final state.
HT, the scalar sum of the transverse momenta of the leptons and jets in the event, for the e±e± (left), e±μ± (center) and μ±μ± final state.
HTmiss, the magnitude of the vector sum of the transverse momenta of the leptons and jets in the event, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Trailing lepton transverse momentum, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Trailing lepton pseudorapidity, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Angular separation between the trailing lepton and the closest hadronic jet, for the e±e± (left), e±μ± (center) and μ±μ± final state.
Transverse mass of the leading lepton – ETmiss system,, for the e±e± (left), e±μ± (center) and μ±μ± final state.

Trilepton final state: events with one lepton failing the lepton MVA

Common caption:These plots show the distribution of the main discriminating variables and the final BDT output for events where one lepton fails the lepton MVA requirement, that are then used to determine the reducible background predictions in the signal region.
In these plots, the expected distribution for the reducible background is taken from simulations, and the yield is fitted from the data. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties after the fit (light blue).

Plot Caption
BDT discriminator output.
Jet multiplicity.
HT, the scalar sum of the transverse momenta of the leptons and jets in the event.
Fraction of HT from objects within |η|<1.2 with respect to the total HT.
pT of the jet with highest b-tagging discriminator.
Maximum |η| among the three leptons.
ΔR between the two closest opposite-sign leptons.
Invariant mass of the reconstructed hadronically-decaying top quark candidate (all events). The bin at zero contains the events for which no hadronically-decaying top quark candidate is reconstructed (events with only two jets, or three jets of which two are b-tagged).
Invariant mass of the reconstructed hadronically-decaying top quark candidate (selected events). Events with no hadronically-decaying top quark candidate are not included in the plot.

Trilepton final state: events with one on-shell Z boson.

Common caption: These plots show the distribution of the main discriminating variables and the final BDT output for events that contain a pair of same-flavour opposite-sign leptons with invariant mass within 10 GeV from that of an on-shell Z boson, but that satisfy all other requirements of the event selection in the trilepton final state.
The background predictions are derived from other control regions and simulations, without fitting the the data in the signal region or in this control region. The bottom panel of each plot shows the ratio between data and predictions with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).

Plot Caption
BDT discriminator output.
Jet multiplicity.
HT, the scalar sum of the transverse momenta of the leptons and jets in the event.
Fraction of HT from objects within |η|<1.2 with respect to the total HT.
pT of the jet with highest b-tagging discriminator.
Maximum |η| among the three leptons.
ΔR between the two closest opposite-sign leptons.
Invariant mass of the reconstructed hadronically-decaying top quark candidate (all events). The bin at zero contains the events for which no hadronically-decaying top quark candidate is reconstructed (events with only two jets, or three jets of which two are b-tagged).
Invariant mass of the reconstructed hadronically-decaying top quark candidate (selected events). Events with no hadronically-decaying top quark candidate are not included in the plot.

Standard candles: WZ → 3l

Common caption: These plots show the distribution of some of the main kinematic variables in fully leptonic WZ events selected using the same tight lepton MVA requirement used for the dilepton and trilepton channel, with the request that there be a pair of same-flavour opposite-sign leptons with invariant mass within 10 GeV of the nominal Z mass. The Z+jets background is suppressed by a tight requirement on the ETmiss linear discriminator, and backgrounds from processes including the top quark are reduced by vetoing events with at least one jet satisfying the medium working point of the CSV b-tagger.
The predictions for WZ and ZZ yields are derived from simulations, and are not fitted to the data. The bottom panel of each plot shows the ratio between data and predictions, with the statistical-only uncertainties (dark blue) and overall uncertainties (light blue).
The purpose of this control region is to validate the modelling of prompt leptons and genuine ETmiss from neutrinos after the selection.

Plot Caption
Transverse momentum of the leading lepton.
Transverse momentum of the trailing lepton.
Transverse mass of the W boson, reconstructed from the lepton not from the Z and the ETmiss.
Missing transverse momentum, ETmiss.

Standard candles: ttZ → 3l

Common caption: These plots show the distribution of some of the main kinematic variables in ttZ events selected using the requirements of the trilepton final state, with the request that there be a pair of same-flavour opposite-sign leptons with invariant mass within 10 GeV of the nominal Z mass, and the request that there be at least 4 hadronic jets.
The predictions for ttZ are derived from simulations, while the ones for the reducible background and WZ are derived from control regions as in the nominal analysis. The bottom panel of each plot shows the ratio between data and predictions, with the statistical-only uncertainties.
The purpose of this control region is to check that the ttZ process can be cleanly identified in a topology where all final state objects are reconstructed (leptons, ETmiss, b-jets from the top quark and light-flavoured jets from the W).

Plot Caption
Invariant mass of the reconstructed hadronically-decaying W boson.
Invariant mass of the two leptons from the Z boson.
Invariant mass of the reconstructed hadronically-decaying top quark.
Multiplicity of hadronic jets.
Multiplicity of b-jets satisfying the medium working point of the CSV tagger.
Transverse momentum of the trailing lepton.

Standard candles: Z → 4l

Common caption: These plots show some kinematic and lepton id distributions for events on the Z → 4l peak, selected using the loose lepton MVA requirements as in the nominal selection for the four-lepton final state.
The prediction for the Z→4l process are derived from simulation, with corrections derived from Z→ll events. The bottom panel of each plot shows the ratio between data and predictions, with the statistical-only uncertainties.
The purpose of this control region is to validate the identification leptons down to to the lowest pT and identification requirements used in the four-lepton final state.

Plot Caption
Invariant mass of the reconstructed four-lepton system.
Transverse momentum of the third lepton (ordered by pT).
Pseudo-rapidity of the third lepton.
Transverse momentum of the fourth lepton (ordered by pT). Note that the binning is not uniform, and the data points and histograms represent a number of events per GeV, not per bin.
Pseudo-rapidity of the fourth lepton.
Lepton MVA discriminator for the less signal-like among the two leptons with largest pT. The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.
Lepton MVA discriminator for the less signal-like among the two leptons with smallest pT. The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.

Data/mc comparisons for lepton input MVA variables (signal-like leptons)

Common caption: These plots show the main lepton MVA variables in event samples dominated by signal-like leptons from Z→ll. In all plots, the predictions from the simulation are normalized to the total event yield in data.

Plot Caption
Lepton MVA discriminator value for the trailing lepton in Z→ee (left) and Z→μμ (right) events.
The distributions are not expected to match precisely as the corrections applied to the simulation only correct the per-lepton efficiency as function of the lepton pT and η, not the shape of the MVA discriminator.
Relative isolation computed from charged hadrons, for the trailing lepton in Z→ee (left) and Z→μμ (right) events.
Relative isolation computed from neutral hadrons and photons, for the trailing lepton in Z→ee (left) and Z→μμ (right) events.
Logarithm of the transverse impact parameter for the trailing lepton in Z→μμ events (this variable is not used in the MVA for electrons).
The MVA uses the logarithm of the impact parameter instead of the impact parameter because it has a more gaussian-like shape, making the training easier.
Logarithm of the longitudinal impact parameter for the trailing lepton in Z→μμ events (this variable is not used in the MVA for electrons).
The MVA uses the logarithm of the impact parameter instead of the impact parameter because it has a more gaussian-like shape, making the training easier.
Largest 3D impact parameter significance for the trailing lepton in Z→ee (left) and Z→μμ (right) events.
Ratio of the lepton pT to the associated jet pT for the trailing lepton in Z→ee (left) and Z→μμ (right) events./b>
b-tagging discriminator of the associated jet for the trailing lepton in Z→ee (left) and Z→μμ (right) events./b>

Data/mc comparisons for lepton input MVA variables (background-like leptons)

Common caption: These plots show the main lepton MVA variables in event samples dominated by background-like leptons. In all plots, the predictions from the simulation are normalized to the total event yield in data.

Plot Caption
Lepton MVA discriminator value for the additional lepton in Z+l (left) or tt+l (right) events.
A precise matching between the data and simulations is not necessary since anyway in the estimation of the reducible background the probability for a non-prompt lepton to satisfy the lepton MVA requirement are taken from data and not from simulations.
Relative isolation computed from charged hadrons, for the additional lepton in Z+l (left) or tt+l (right) events.
Relative isolation computed from neutral hadrons and photons, for the additional lepton in Z+l (left) or tt+l (right) events.
Logarithm of the transverse impact parameter for the additional lepton in Z+l (left) or tt+l (right) events.
The MVA uses the logarithm of the impact parameter instead of the impact parameter because it has a more gaussian-like shape, making the training easier.
Logarithm of the longitudinal impact parameter for the additional lepton in Z+l (left) or tt+l (right) events.
The MVA uses the logarithm of the impact parameter instead of the impact parameter because it has a more gaussian-like shape, making the training easier.
Largest 3D impact parameter significance for the additional lepton in Z+l (left) or tt+l (right) events.
Ratio of the lepton pT to the associated jet pT for the additional lepton in Z+l (left) or tt+l (right) events
b-tagging discriminator of the associated jet for the additional lepton in Z+l (left) or tt+l (right) events

Feynman diagrams

Common caption: Examples of leading-order Feynman diagrams for ttH production at pp colliders, followed by Higgs boson decays to ττ, ZZ* and WW* in the signatures considered in this analysis.

Plot Caption
same-sign dilepton final state from a H→WW decay
trilepton final state from a H→WW→2l2ν decay
trilepton final state from a H→WW→lνqq; decay
four-lepton final state from a H→WW decay
same-sign dilepton final state from a H→ττ decay
trilepton final state from a H→ZZ decay
Edit | Attach | Watch | Print version | History: r16 < r15 < r14 < r13 < r12 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r16 - 2020-10-22 - GiovanniPetrucciani
 
    • 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 & 2008-2021 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