Search for supersymmetry in events with opposite-sign dileptons and missing energy using an ANN

Abstract : We present a search for supersymmetry (SUSY) in events with two opposite sign isolated leptons in the final state, accompanied by hadronic jets and missing transverse energy. Advanced multivariate techniques, and in particular Artifical Neural Networks, are deployed in order to discriminate between possible SUSY signals from the Standard Model backgrounds. The analysis uses a data sample collected with the CMS detector during the 2011 LHC run and corresponding to an integrated luminosity of 2.3/fb. This search is independent and complementary to the ones already published so far, being more inclusive for models which exhibit low MET/ low HT (MET>30 GeV and HT>80 GeV) event characteristics. Good agreement is observed between expectation and observation, yielding limits in the SMS (Simplified Models) and cMSSM planes.

Link to documentation in CDS


Figure Abbreviated Caption

MET for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Primary Jet pT for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Secondary Jet pT for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Number of Jets for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Transverse event mass for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Invariant di-lepton mass for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Ratio of lepton ET to total ET for simulated SM backgrounds (red), CMSSM LM6 benchmark point (blue), and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Control region to signal Region extrapolation factor as obtained from simulated events.
logo ANN output for the SM background (red line) and CMSSM LM6 benchmark point (blue line)
logo Efficiency (red) and purity (blue) vs ANN output cut for CMSSM LM6 benchmark point, after the event and preselection criteria are applied.
logo Data (black points) vs MC (red lines) comparisons of the ANN output in the control region used in order to perform the extrapolation with 20% systematic uncertainties included.
logo Ratio between data and simulated events of the ANN output in the control region, with both statistical (black error bars) and systematic (red bands) uncertainties shown
logo Data (black points) vs MC (red lines) comparisons of the ANN output distributions in the ttbar enriched control region
logo Data (black points) vs MC (red lines) comparisons of the ANN output distributions in the Z+jets enriched control region
logo ANN output for the data (black points) and the total SM data-driven prediction (gray hatched area) with systematic uncertainties in the signal region. The various background contributions to the total SM prediction are shown in colour.
logo 95% C.L. exclusion limits on Simplified Models of the inclusive ANN analysis (ANN> 0.8).
logo Expected (blue) and observed (red) 95% C.L. exclusion limit for the ANN inclusive analysis (ANN> 0.8) for the CMSSM plane. The 95% C.L. upper limits are computed using a frequentist CLs method with profile likelihood test statistics, and truncated Gaussian distribution for the background expectation. The uncertainties in the NLO cross sections from the parton distribution functions, the choice of the factorization and renormalization scale, and alphaS are taken into account, and are evaluated according to the PDF4LHC recommendation. A signal acceptance systematic of 17% is used for each point.
logo ANN architecture with the seven input variables
logo Error evolution in the cost function minimization procedure for the training (red) and test (blue) populations
logo ANN output for the SM background (red) and CMSSM LM0 benchmark point (blue)
logo ANN output for the SM background (red) and CMSSM LM1 benchmark point (blue)
logo ANN output for the SM background (red) and CMSSM LM9 benchmark point (blue)
logo ANN output for the SM background (red) and GGM (tanbeta=10) sample (blue)
logo ANN output for the SM background (red) and GGM (tanbeta=20) sample (blue)
logo ANN output for the SM background (red) and GMSM sample (blue)

logo MET for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Primary jet pT for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Secondary jet pT for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Number of jets for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Transverse event mass for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Invariant dilepton mass for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Ratio of lepton ET to total ET for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo Dilepton pair type for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the control region. Statistical and systematic errors on the data and simulated events are shown
logo MET for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Primary jet pT for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Secondary jet pT for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Number of jets for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Transverse event mass for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Invariant dilepton mass for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Ratio of lepton ET to total ET for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown
logo Dilepton pair type for simulated SM backgrounds (red) and data (black points) normalized to the same number of events in the signal region. Statistical and systematic errors on the data and simulated events are shown

-- NikiSaoulidou - 04-Mar-2012

Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r3 - 2012-03-11 - EvaHalkiadakis
 
    • 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