IFAEAtlasAnalysis

Introduction

This page is created by IFAE Barcelona. It is supposed to collect or share useful information about ATLAS analysis. In order to edit TWiki pages you have to register with your AFS password

Useful links

• IFAE Cvs Repository IFAECvsRepo
• CSC data
  • Status of AOD replicas on sites
  • The simple and fast overview - highly recommended
  • List of datasets being reprocessed due to the 30 microns bug
  • panda monitor
  • Atlas Metadata Interface
  • Requested data sets
  • DC3 requests
  • Naming convention
• Grid
  • DQ2 manual with trouble shooting
• Analysis tools
  • Introduction into the Atlas analysis model
  • Introduction into HighPtView
    • How to run Highptview
  • Introduction into TopView
  • Introduction into EventView
    • Tutorial about configurables: EventViewBuilderTutorialConf
    • Our own collected information: IFAEAtlasAnalysisEV
  • Introduction into StructuredAthenaAwareNtuple
• MC production
  • Introduction: see Running Athena in the Atlas WorkBook
  • MCFM main page
    • LHADPF: Les Houches library
• General information
  • Hypernews forum

Trigger

• official pages
  • Trigger Hypothesis
  • Trigger Analysis Tutorial (updated)
  • Trigger Menu

• General information
  • Trigger performance overview by Monica Wiehlers in the Exotics meeting

• detailed info
  • Egamma triggers
  • Egamma trigger performance optimization, by Iwona Grabowska-Bold

Z+jets

Object ID

Electron Reconstruction

• Short description and Documentation
  • I'm starting documentation in EventViewElectronDetailed, but this will take some time..
  • As long as the CSC note is not finished, the best EMID documentation can be found in the EgammaATLASRecoPerformance page. Here you can find the link to the most recent performance note.
  • A complete list of the AOD content can be accessed from the Class references in the AOD pages
  • A project to document the AOD content has started (Kevin Black). So far the mosy useful documentation is the EgammaCBNT page and the documentation of the most important Electron Variables produced by the EV calculators in 12.0.x
  • Electron candidates are separated in reconstruction from Photon candidates by applying a loose track match. The default electron ID uses a hollow cone isolation and the IsEM flag which is composed of bits that correspond to various requirements on the shower shape in the calorimeter and the track in the ID (see ElectronGammaIsEM page). Alternative approaches using multivariate techniques (NN, Likelihood, HMatrix, ..) are in advanced stage of development.
  • A good source of information are also various recent talks in the EMID group
    • Egamma Status by S. Rajagopalan on the Stockholm ATLAS Overview Week
    • Comparative Electron ID studies by Jamie Boyd
    • A status report of Christophe Clement in the CAT meeting which contains documentations and distributions of the ID variables for electrons and fakes and the corresponding presentation in the T and P week.
    • A presentation of Gia Khoriauli about EM ID in 12.0.4
    • EM calibration in 12.0.4 by Claire Adam Bourdarios
• More details about the Electron ID
  • AOD electron :
    • Sliding Window Clusters in 5x5, 3x5 and 3x7
    • Clusters corrected for position and energy modulations
    • Separation of Electron and Photon candidates by a simple track match: DEta<0.025 , DPhi<0.05; E/p<4
    • Hadronic activitity < 20%
  • Is EM is calculated (no cut applied in AOD), using information about the Hadronic Leakage, the lateral shower shape in the second sampling of the EM Calorimeter, the lateral shower shape in the strips, the track Quality Cuts (# of hits and impact parameter), the track-calorimeter spatial matching, E/p and the fraction of high threshold TRT hits
    • More information: isEM bit mask
  • The electron likelihood parametrizes the distributions of various discriminating variables for signal (S) and background (B) which are then interpreted as PDFs. The signal probability S ( or em Weight) and the background probability B (or pion Weight) are calculated as a product of the individual probabilities. The final discriminant is the likelihood ratio S/(S+B). The ATLAS electron likelihod uses e0/e, e1/e, e2/e, e3/e, Et Cone40/et, E237/E277, E233/E237,et/(et+Etha), e/p, deta1, dphi2, zvertex/errz, Weta1, Weta2, E2tsts1-Emins1, Fracs1, Isol and numtracks (see documentation of vaiables )

Muon Reconstruction

• Documentation
  • Combined Muon Status by Peter Kluit at the Stockholm ATLAS Overview Week
  • MuonRecoPedia information about Atlas Muon reconstrucion software.
  • Muon Selection & Performance by David Adams at th U.S. muon software workshop (nice talk) (6/12/2006)
  • Muon Performance in 12.0.X by David Adams at the Muon Validation & Performance meeting (23/01/2007)
  • EV variables: -> muon part of the EVUD calculator documentation

Jet Reconstruction

• Documentation
  • Jet reconstruction
  • Jet-Tau-ETmiss Status Report bt Donatella Cavalli at the Stockholm ATLAS Overview Week
  • EV variables: -> jet part of the EVUD calculator documentation

MC data sets

• Streaming test:
  • wiki page for users

• Pythia Z-> leplep inclusive
  • Documentation
  • Z->ee: 5144 trig1_misal1_csc11.005144.PythiaZee.recon.AOD.v12000601
  • Z->mumu: 5145 trig1_misal1_csc11.005145.PythiaZmumu.recon.AOD.v12000601
  • Z->tautau->lep: 5146 trig1_misal1_csc11.005146.PythiaZtautau.recon.AOD.v12000601
• Higgs group Z+jets Loose samples:
  • Information:
    • generation conditions: ZplusJetsAlpgen
    • data sets: StandardModelCSCDatasets
  • ID 8130 .... 8135 (0 jets .... >=5 jets)
  • atlfast samples: /castor/cern.ch/grid/atlas/users/jtanaka/HiggsWG/Z_samples/fast/CBNT/
  • fully reconstructed Zee: trig1_misal1_mc12.008131.AlpgenJimmyZeeNp1LooseCut.recon.AOD.v12000601 etc.
  • fully reconstructed Zmm: trig1_misal1_mc12.008143.AlpgenJimmyZmumuNp1LooseCut.recon.AOD.v12000601 etc
• Alpgen Susy background: Z->nunu
  • Information: ZplusJetsAlpgen
  • ID: 8124 ... 8129
  • atlfast samples (directory contains also other SUSY background ): /castor/cern.ch/grid/atlas/datafiles/susy/atlfast/2006_b
• QCD background
  • for Zmm: trig1_misal1_mc12.017506.PythiaB_bbmu6mu4X.recon.AOD.v12000601
  • for Zee: trig1_misal1_mc12.005802.JF17_pythia_jet_filter.recon.AOD.v12000601
• Pythia SUSY background
  • Overview and production status

Zmumu+jets analysis

• workplans:
  • workplan16Gener.pdf

TTbar-> tau

MC samples

• Top MC samples are defined here. Any sample reconstructed with release 11 can found at panda monitor. Production with release 12 is on going, not much was produced until now (for Top analysis), however some few AODs were reconstructed by Elzbieta, and can be found at:
  • W->e nu: 5104 = /castor/cern.ch/grid/atlas/users/ashibata/cscpre12-5104
  • Z->tautau: 5188 = /castor/cern.ch/grid/atlas/users/ashibata/cscpre12-5200
  • ttbar: 5200 = /castor/cern.ch/grid/atlas/users/ashibata/cscpre12-5200
  • single top: 5501 = /castor/cern.ch/grid/atlas/users/ashibata/cscpre12-5200
• And replicated at pic at:
  • /nfs/atlas-data03/cosuna/official_data/cscpre12/csc11.005200.T1_McAtNlo_Jimmy.digit.RDO.v11004205
• As the statistics for rel12 is not enough, we reconstructed more events (running on rel11 digit files) and placed them under:
  • /nfs/atlas-data03/cosuna/official_data/cscpre12/csc11.005200.T1_McAtNlo_Jimmy.digit.RDO.v11004205_IFAE
• Ntuples produced with the leptonic W's daughter-number fix are placed under:
  • /nfs/atlas-data03/fiorini/cspre12/ntuples TopViewAANtuple.new.root has been obtained merging the 2 other files.

TopView ntuples

• Ongoing detailed documentation on ntuple variables.
• For a full description of ntuple variables: Truth, Reco. Another description (should be the "official") here.
• release 11. Large TopView athena aware ntuples were produced with release 11.
• release 12. Few files were produced for release 12 (and TopView-00-12-xx ) for validation. With this release two types of AAN are being generated, depending on the type of tau identification algorithm used (tauRec or tau1p3p):
  • /castor/cern.ch/grid/atlas/users/ashibata/cscpre12-TopView.
• Ntuples used in the analysis and corresponding MC info.

B-tagging

B tagging is done by default combining the results of 2 different algorithms: IP3D and SV1.

• IP3D calculates with tracks informations the 3-D impact parameter (d) of a b vertex candidate and its significance: s = d/sigma(d).
• SV1 reconstructs secondary vertices in the BJet. SV1 algorithm uses a bottom-up approach: it fits vertices with all pairs of tracks, retaining the fit with the highest probability and fitting the remaining tracks of the jet to this vertex. All tracks below a certain fit probability will be rejected. A likelihood variable is calculated with the secondary vertex parameters (mass, fit probability, multiplicity and distance from the reconstructed primary vertex).
• The two informations are summed in a bjet weight variable, that is variable used to discriminate bjets from other jets. The standard cut on the weight is at value of 3 (b-tagged jets having a weight greater than 3). The cut can be changed in TopView inside python/TopViewInserters_module.py . The default tagging algorithm has an efficiency of 60% and a Rejection factor for light jets of 240 in ttjjenu and ttH events.
• More details are available at BTagging Homepage and the main BTagging algorithm is in CVS under: PhysicsAnalysis/JetTagging/JetTagAlgs/BTagging
• ...

SUSY

Links

• Atlas Susy working group
• Susy MC production status
• Susy MC production information
• Old page for Rome production

Data

• Monitoring of Susy CSC production
• Cross sections for alpgen samples

EV groupArea at PIC.

The PAT group is continuosly developing and applying patched to official atlas releases, even when the release is out. Therefore they make use the so called groupArea, where they install updates packages that will overload the default packages to be found in the release. In order to use the latest EV, SUSYView, TopView,... we need to install a given EVTag at PIC (exactly in the same way as they do in groupArea at CERN). There are several ways one can do that (consult here), but the easiest and faster is to use pacman.

I installed the latest EV version at:

/nfs/pic.es/user/c/cosuna/scratch/public/EVgroupArea/EventView-12.0.4.3

In order to use this area, prepend this path to your CMTPATH before running athena:

export CMTPATH=/nfs/pic.es/user/c/cosuna/scratch/public/EVgroupArea/EventView-12.0.4.3:$CMTPATH

Future updates can be installed in the same EVgroupArea directory using pacman

• Using Pacman at PIC

Let say you want to install EventView-$Release at PIC. First configure pacman:

      * cd /nfs/pic.es/user/c/cosuna/scratch/public/pacman/pacman-$VERSION
      * source setup.sh

Then go to the EV groupArea directory and create a new directory for the desired version

      * cd /nfs/pic.es/user/c/cosuna/scratch/public/EVgroupArea/
      * mkdir EventView-$RELEASE
      * cd EventView-$RELEASE

and execute the pacman command:

      * pacman -v all -get http://atlas-computing.web.cern.ch/atlas-computing/links/kitsDirectory/PAT/EventView/:EventView-$RELEASE.pacman

Bibliography

Under this section are stored some interesting documents, talks and lectures that you can find helpful to consult during your analysis:

• K. Ellis Introduction to QCD at Colliders

---

Major updates:
-- UllaBlumenschein - 09 Jan 2007 -- Main.cosuna - 09 Jan 2007

%RESPONSIBLE% Main.unknown

%REVIEW% Never reviewed