Photon Energy Regression Studies at IPNL

Code :

Traditionnal Regression

This code trains the regression, then tests it producing some illustrative plots

scram project CMSSW_7_0_1
cd CMSSW_7_0_1/src/
cmsenv
git clone https://github.com/bcourbon/TraditionalRegression.git
cd TraditionalRegression/
source /afs/cern.ch/sw/lcg/app/releases/ROOT/5.34.04/x86_64-slc5-gcc46-opt/root/bin/thisroot.sh
make
mkdir plots
./regression.exe run/MyGBR_Photons_70pre11.config

The input file is specified in MyGBR_Photons_70pre11.config

Don't forget to install eos; to install eos :

cd 
mkdir eos
eosmount eos

To produce the input trees by yourself :

cd CMSSW_7_0_1/src/
cmsenv
git clone https://github.com/camilocarrillo/GedPhotonRegTreeMaker.git
cd GedPhotonRegTreeMaker/GedPhotonRegTreeMaker/ 
scram b 
cd python/
cmsRun runPhotonRegressionTrees_cfg.py

Semi-Parametric Regression

This code trains the regression, then tests it producing some illustrative plots

scram project CMSSW_6_2_x
cd CMSSW_6_2_x/src/
cmsenv
git clone https://github.com/bcourbon/SemiParametricRegression.git
cd SemiParametricRegression/
scram b
root -l macros/RegressionTraining.C+\(true\) #barrel
root -l macros/RegressionTraining.C+\(false\) #endcaps
root -l macros/RegressionTesting.C+\(true\) #barrel
root -l macros/RegressionTesting.C+\(false\) #endcaps
hadd -f resultsEBEE.root resultsEB.root resultsEE.root
mkdir plots
root -l macros/RegressionPlotting.C+
root -l macros/RegressionPlotting.C+\(false,true,false\) #to plot results only for barrel
root -l macros/RegressionPlotting.C+\(false,false,true\) #to plot results only for endcaps

Don't forget to install eos if you want to use the input files specified in the code

Documentation :

• Presentations from Jean-Baptiste
  • https://indico.cern.ch/event/244507/session/12/contribution/21/material/slides/0.pdf
  • https://indico.cern.ch/event/241450/contribution/2/material/slides/0.pdf
  • https://indico.cern.ch/event/226227/contribution/0/material/slides/0.pdf
  • https://indico.cern.ch/event/240259/contribution/1/material/slides/0.pdf
  • https://indico.cern.ch/event/278380/contribution/0/material/slides/0.pdf
  • https://indico.cern.ch/event/244507/session/19/contribution/42/material/slides/0.pdf
  • https://indico.cern.ch/event/293993/contribution/3/material/slides/0.pdf
  • https://indico.cern.ch/event/219368/contribution/3/material/slides/0.pdf

• Presentations from Josh
  • https://indico.cern.ch/event/270535/contribution/1/material/slides/0.pdf
  • https://indico.cern.ch/event/152992/contribution/5/material/slides/0.pdf
  • https://indico.cern.ch/event/155683/session/6/contribution/14/material/slides/0.pdf
  • https://indico.cern.ch/event/189919/session/3/contribution/52/material/slides/0.pdf
  • https://indico.cern.ch/event/306030/contribution/4/material/slides/0.pdf

• Hgg Analysis note (section 6 : Photon energy reconstruction)
  • https://twiki.cern.ch/twiki/pub/CMS/HggAnalyisNote13253/AN-13-253_temp.pdf

• Friedman paper on machine learning
  • http://statweb.stanford.edu/~jhf/ftp/trebst.pdf

-- CamiloCarrillo - 06 May 2014