wiki:PublicAnalysisDatabase

Version 80 (modified by kraml, 5 years ago) (diff)

--

MadAnalysis 5 physics analysis database

This page contains a collection of LHC analyses that have been implemented in the MadAnalysis 5 framework (in the expert mode) described in arXiv:1405.3982] and arXiv:1407.3278]. For each analysis, a commented C++ code is available for download, provided together with information on the analysis and references to relevant publications. A careful validation of each implementation is also provided.

In order to use these files, you need MadAnalysis 5 version 1.1.12, which can be downloaded here (beware, the latest version on Launchpad is 1.1.11 and will not have all the functionality described belo). After installation of MadAnalysis 5 v1.1.12, you also need our modified version of Delphes. To install it, type

 install delphesMA5tune

within MadAnalysis 5. If you have an older version of MadAnalysis 5, first update it to the latest one. Also, beware that the standard Delphes and delphesMA5tune cannot be used simultaneously!!

In order to use the code, you first need to create a working directory (that will be called PAD). You can also optionally install a couple of python tools that allows for limit and efficiency map extractions. However, in order to have those tools available in the PAD directory, they need to be installed prior to the creation of the working directory. You need to start MadAnalysis 5 as

 ./bin/ma5 -R

If you want to use our scripts for limit setting and efficiency maps (see below), next type

 install RecastingTools

For creating a working directory that contains all the analyses that are publicly available, then type

install PAD

New analyses (within the PAD or separately) can be added by running the script newAnalyzer.py (located in Build/SampleAnalyzer), passing as an argument the name of the analysis of interest. For instance,

 python newAnalyzer.py cms_sus_14_001

will create and modify the necessary files for adding the analysis named cms_sus_14_001. In particular, a pair of header and source C++ files cms_sus_14_001.cpp and cms_sus_14_001.h are now available in the subdirectory Build/SampleAnalyzer/User/Analyzer. You can now either implement the analysis yourself or replace the newly created files with those shared by somebody else (or downloaded from the list below in case you do not use the automatic PAD installation from above).

After execution of an analysis, exclusion under the CLs prescription can be computed with the Python code exclusion_CLs.py Note that it requires SciPy to be installed. The path to the working directory of interest has to be provided in the variable analysis_path in the beginning of the code. If you have done install RecastingTools, then a local copy of exclusion_CLs.py is available in every new working directory. The code is called as

 ./exclusion_CLs.py analysis_name mypoint.txt [run_number] [cross section in pb]

The exclusion results will be printed on the screen and in an output file located in the directory mypoint.txt/ of the output of the analysis. Details are given in arXiv:1407.3278.

In order to get the efficiency map associated with a given benchmark point, the getmaps.py script can be run

 ./getmaps.py mypoint.txt [cross section value]

where [cross section value] is the value of the cross section associated to the generated event sample under consideration.This creates one file for each analysis that has been executed, this file containing the efficiency associated with each of the signal regions of the relevant analysis. The efficiency maps come with the computed relative (statistical errors). In the case the script is run with a cross section value, the best expected region is indicated in the output file. (We stress that in order to have these scripts available, they must be installed before creating the working directory.) In the case the cross section value is negative, the best expected region derivation is omitted and only the efficiency map is computed.

We strongly encourage users to publish their re-implemented LHC analyses via Inspire. This way they not only become part of the public database, rendering them as useful as possible to the community, but they also become individually searchable and citable.

For any information on the implementation of an analysis, on the database or on the expert mode of the program, please visit the 'Answers' section of our launchpad page.

Available Analyses

!! please properly cite all the re-implementation codes you are using (see Inspire citation entry) !!

ATLAS analyses, 8 TeV

Analysis Short Description Implemented by Code Validation note Status
ATLAS-SUSY-2013-05 (published) stop/sbottom search: 0 leptons + 2 b-jets G. Chalons Inspire PDF (figures) done
ATLAS-SUSY-2013-11 (published) EWK-inos, 2 leptons + MET B. Dumont Inspire PDF (source) done
ATLAS-HIGG-2013-03 (published) ZH->ll+invisible B. Dumont Inspire PDF (source) done
ATLAS-EXOT-2014-06 (published) mono-photons + MET D. Barducci Inspire PDF MadGraph cards done

Delphes card for ATLAS-SUSY-2013-05 and ​ATLAS-EXOT-2014-06
Delphes card for ATLAS-SUSY-2013-11 and ATLAS-HIGG-2013-03

CMS analyses, 8 TeV

Analysis Short Description Implemented by Code Validation note Status
CMS-SUS-13-011 (published) stop search in the single lepton mode B. Dumont, B. Fuks, C. Wymant Inspire [1]
PDF (source) done
CMS-SUS-13-012 (published) gluino/squark search in jet multiplicity and missing energy S. Bein, D. Sengupta Inspire
PDF (source) done
CMS-SUS-13-016 (PAS) search for gluinos using OS dileptons and b-jets D. Sengupta, S. Kulkarni Inspire PDF(source) done

Delphes card for these analyses

[1] This analysis requires MINUIT libraries. Therefore, the line

 LIBFLAGS += -lMinuit

should be added to the Makefile of the Build/ directory before compilation.

Attachments (77)