// main GUI
void TMVAGui( const char* fName = "TMVA.root" )
{
// Use this script in order to run the various individual macros
// that plot the output of TMVA (e.g. running TMVAnalysis.C),
// stored in the file "TMVA.root"
// for further documentation, look in the individual macros
TString curMacroPath(gROOT->GetMacroPath());
gROOT->SetMacroPath(curMacroPath+":$ROOTSYS/tmva/test/:");
cout << "--- Launch TMVA GUI to view input file: " << fName << endl;
// init
TMVAGui_inactiveButtons.clear();
// check if file exist
TFile* file = TFile::Open( fName );
if (!file) {
cout << "==> Abort TMVAGui, please verify filename" << endl;
return;
}
// find all references
cout << "--- Reading keys ..." << endl;
TMVAGui_keyContent = (TList*)file->GetListOfKeys()->Clone();
// close file
file->Close();
TString defaultRequiredClassifier = "";
// gROOT->Reset();
// gStyle->SetScreenFactor(2); // if you have a large screen, select 1,2 or 1.4
// create the control bar
TControlBar* cbar = new TControlBar( "vertical", "TMVA Plotting Macros", 0, 0 );
const TString buttonType( "button" );
// configure buttons
Int_t ic = 0;
ActionButton( cbar,
Form( "(%ia) Input Variables (training sample)", ++ic),
Form( ".x variables.C(\"%s\",0)", fName ),
"Plots all input variables (macro variables.C)",
buttonType );
ActionButton( cbar,
Form( "(%ib) Decorrelated Input Variables", ic ),
Form( ".x variables.C(\"%s\",1)", fName ),
"Plots all decorrelated input variables (macro variables.C(1))",
buttonType, "DecorrTransform" );
ActionButton( cbar,
Form( "(%ic) PCA-transformed Input Variables", ic ),
Form( ".x variables.C(\"%s\",2)", fName ),
"Plots all PCA-transformed input variables (macro variables.C(2))",
buttonType, "PCATransform" );
ActionButton( cbar,
Form( "(%id) GaussDecorr-transformed Input Variables", ic ),
Form( ".x variables.C(\"%s\",3)", fName ),
"Plots all GaussDecorrelated-transformed input variables (macro variables.C(3))",
buttonType, "GaussDecorr" );
ActionButton( cbar,
Form( "(%ia) Input Variable Correlations (scatter profiles)", ++ic ),
Form( ".x CorrGui.C\(\"%s\",0)", fName ),
"Plots signal and background correlation profiles between input variables (macro CorrGui.C)",
buttonType );
ActionButton( cbar,
Form( "(%ib) Decorrelated Input Variable Correlations (scatter profiles)", ic ),
Form( ".x CorrGui.C\(\"%s\",1)", fName ),
"Plots signal and background correlation profiles between decorrelated input variables (macro CorrGui.C(1))",
buttonType, "DecorrTransform" );
ActionButton( cbar,
Form( "(%ic) PCA-transformed Input Variable Correlations (scatter profiles)", ic ),
Form( ".x CorrGui.C\(\"%s\",2)", fName ),
"Plots signal and background correlation profiles between PCA-transformed input variables (macro CorrGui.C(2))",
buttonType, "PCATransform" );
ActionButton( cbar,
Form( "(%id) GaussDecorr-transformed Input Variable Correlations (scatter profiles)", ic ),
Form( ".x CorrGui.C\(\"%s\",3)", fName ),
"Plots signal and background correlation profiles between Gaussianised and Decorrelated input variables (macro CorrGui.C(3))",
buttonType, "GaussDecorr" );
ActionButton( cbar,
Form( "(%i) Input Variable Linear Correlation Coefficients", ++ic ),
Form( ".x correlations.C(\"%s\")", fName ),
"Plots signal and background correlation summaries for all input variables (macro correlations.C)",
buttonType );
ActionButton( cbar,
Form( "(%ia) Classifier Output Distributions (test sample)", ++ic ),
Form( ".x mvas.C(\"%s\",0)", fName ),
"Plots the output of each classifier for the test data (macro mvas.C(...,0))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ib) Classifier Output Distributions for Training and Test Samples", ic ),
Form( ".x mvas.C(\"%s\",3)", fName ),
"Plots the rarity of each classifier for the test data (macro mvas.C(...,3))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ic) Classifier Probability Distributions", ic ),
Form( ".x mvas.C(\"%s\",1)", fName ),
"Plots the probability of each classifier for the test data (macro mvas.C(...,1))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%id) Classifier Rarity Distributions", ic ),
Form( ".x mvas.C(\"%s\",2)", fName ),
"Plots the rarity of each classifier for the test data (macro mvas.C(...,2))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Cut Efficiencies", ++ic ),
Form( ".x mvaeffs.C+(\"%s\")", fName ),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ib) Classifier Background Rejection vs Signal Efficiency (ROC curve)", ic ),
Form( ".x efficiencies.C(\"%s\")", fName ),
"Plots background rejection vs signal efficiencies (macro efficiencies.C) [\"ROC\" stands for \"Receiver Operation Characteristics\"]",
buttonType, defaultRequiredClassifier );
TString title = Form( "(%i) Parallel Coordinates (requires ROOT-version >= 5.17)", ++ic );
ActionButton( cbar,
title,
Form( ".x paracoor.C(\"%s\")", fName ),
"Plots parallel coordinates for classifiers and input variables (macro paracoor.C, requires ROOT >= 5.17)",
buttonType, defaultRequiredClassifier );
// parallel coordinates only exist since ROOT 5.17
#if ROOT_VERSION_CODE < ROOT_VERSION(5,17,0)
TMVAGui_inactiveButtons.push_back( title );
#endif
ActionButton( cbar,
Form( "(%i) Likelihood Reference Distributiuons", ++ic),
Form( ".x likelihoodrefs.C(\"%s\")", fName ),
"Plots to verify the likelihood reference distributions (macro likelihoodrefs.C)",
buttonType, "Likelihood" );
ActionButton( cbar,
Form( "(%ia) Network Architecture", ++ic ),
Form( ".x network.C(\"%s\")", fName ),
"Plots the MLP weights (macro network.C)",
buttonType, "MLP" );
ActionButton( cbar,
Form( "(%ib) Network Convergence Test", ic ),
Form( ".x annconvergencetest.C(\"%s\")", fName ),
"Plots error estimator versus training epoch for training and test samples (macro annconvergencetest.C)",
buttonType, "MLP" );
ActionButton( cbar,
Form( "(%i) Decision Trees", ++ic ),
Form( ".x BDT.C+(\"%s\")", fName ),
"Plots the Decision Trees trained by BDT algorithms (macro BDT.C(itree,...))",
buttonType, "BDT" );
ActionButton( cbar,
Form( "(%i) Decision Tree Control Plots", ++ic ),
Form( ".x BDTControlPlots.C(\"%s\")", fName ),
"Plots to monitor boosting and pruning of decision trees (macro BDTControlPlots.C)",
buttonType, "BDT" );
ActionButton( cbar,
Form( "(%i) PDFs of Classifiers", ++ic ),
Form( ".x probas.C(\"%s\")", fName ),
"Plots the Fit of the Methods outputs; to plot other trees (i) call macro from command line (macro probas.C(i))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%i) Rule Ensemble Importance Plots", ++ic ),
Form( ".x rulevis.C(\"%s\",0)", fName ),
"Plots all input variables with rule ensemble weights, including linear terms (macro rulevis.C)",
buttonType, "RuleFit" );
cbar->AddSeparator();
cbar->AddButton( Form( "(%i) Quit", ++ic ), ".q", "Quit", buttonType );
// set the style
cbar->SetTextColor("black");
// there seems to be a bug in ROOT: font jumps back to default after pressing on >2 different buttons
// cbar->SetFont("-adobe-helvetica-bold-r-*-*-12-*-*-*-*-*-iso8859-1");
// draw
cbar->Show();
// indicate inactive buttons
for (UInt_t i=0; i<TMVAGui_inactiveButtons.size(); i++) cbar->SetButtonState( TMVAGui_inactiveButtons[i], 3 );
if (TMVAGui_inactiveButtons.size() > 0) {
cout << "=== Note: inactive buttons indicate that the corresponding classifiers were not trained ===" << endl;
}
gROOT->SaveContext();
}
// main GUI
void TMVAGui( const char* fName = "TMVA.root" ) ///usage:
{
//TMVA::gConfig().GetVariablePlotting().fNbins1D = 50.0;
//TMVA::gConfig().GetVariablePlotting().fNbinsXOfROCCurve=50;
// Use this script in order to run the various individual macros
// that plot the output of TMVA (e.g. running TMVAClassification.C),
// stored in the file "TMVA.root"
TString curMacroPath(gROOT->GetMacroPath());
// uncomment next line for macros submitted to next root version
gROOT->SetMacroPath(curMacroPath+":$ROOTSYS/tmva/test/:");
// for the sourceforge version, including $ROOTSYS/tmva/test in the
// macro path is a mistake, especially if "./" was not part of path
// add ../macros to the path (comment out next line for the ROOT version of TMVA)
// gROOT->SetMacroPath(curMacroPath+":../macros:");
cout << "--- Launch TMVA GUI to view input file: " << fName << endl;
// init
TMVAGui_inactiveButtons.clear();
// check if file exist
TFile* file = TFile::Open( fName );
if (!file) {
cout << "==> Abort TMVAGui, please verify filename" << endl;
return;
}
// find all references
cout << "--- Reading keys ..." << endl;
TMVAGui_keyContent = (TList*)file->GetListOfKeys()->Clone();
// close file
file->Close();
TString defaultRequiredClassifier = "";
// gROOT->Reset();
// gStyle->SetScreenFactor(2); // if you have a large screen, select 1,2 or 1.4
// create the control bar
TControlBar* cbar = new TControlBar( "vertical", "TMVA Plotting Macros for Classification", 0, 0 );
const TString buttonType( "button" );
// configure buttons
Int_t ic = 1;
// find all input variables types
TList* keylist = GetKeyList( "InputVariables" );
TListIter it( keylist );
TObjString* str = 0;
char ch = 'a';
while ((str = (TObjString*)it())) {
TString tmp = str->GetString();
TString title = Form( "Input variables '%s'-transformed (training sample)",
tmp.ReplaceAll("InputVariables_","").Data() );
if (tmp.Contains( "Id" )) title = "Input variables (training sample)";
ActionButton( cbar,
Form( "(%i%c) %s", ic, ch++, title.Data() ),
Form( ".x variables.C(\"%s\",\"%s\",\"%s\")", fName, str->GetString().Data(), title.Data() ),
Form( "Plots all '%s'-transformed input variables (macro variables.C(...))", str->GetString().Data() ),
buttonType, str->GetString() );
}
ic++;
// correlation scatter plots
it.Reset(); ch = 'a';
while ((str = (TObjString*)it())) {
TString tmp = str->GetString();
TString title = Form( "Input variable correlations '%s'-transformed (scatter profiles)",
tmp.ReplaceAll("InputVariables_","").Data() );
if (tmp.Contains( "Id" )) title = "Input variable correlations (scatter profiles)";
ActionButton( cbar,
Form( "(%i%c) %s", ic, ch++, title.Data() ),
Form( ".x CorrGui.C(\"%s\",\"%s\",\"%s\")", fName, str->GetString().Data(), title.Data() ),
Form( "Plots all correlation profiles between '%s'-transformed input variables (macro CorrGui.C(...))",
str->GetString().Data() ),
buttonType, str->GetString() );
}
// coefficients
ActionButton( cbar,
Form( "(%i) Input Variable Linear Correlation Coefficients", ++ic ),
Form( ".x correlations.C(\"%s\")", fName ),
"Plots signal and background correlation summaries for all input variables (macro correlations.C)",
buttonType );
ActionButton( cbar,
Form( "(%ia) Classifier Output Distributions (test sample)", ++ic ),
Form( ".x mvas.C(\"%s\",0)", fName ),
"Plots the output of each classifier for the test data (macro mvas.C(...,0))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( " (%ib) Classifier Output Distributions (test and training samples superimposed) ", ic ),
Form( ".x mvas.C(\"%s\",3)", fName ),
"Plots the output of each classifier for the test (histograms) and training (dots) data (macro mvas.C(...,3))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ic) Classifier Probability Distributions (test sample)", ic ),
Form( ".x mvas.C(\"%s\",1)", fName ),
"Plots the probability of each classifier for the test data (macro mvas.C(...,1))",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%id) Classifier Rarity Distributions (test sample)", ic ),
Form( ".x mvas.C(\"%s\",2)", fName ),
"Plots the Rarity of each classifier for the test data (macro mvas.C(...,2)) - background distribution should be uniform",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Cut Efficiencies: S/sqrt(S+B)", ++ic ),
Form( ".x mvaeffs.C+(\"%s\",\"S/sqrt(S+B)\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/sqrt(S+B)", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/sqrt(S+B)\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/sqrt(S+B) @B_Purity=0.95", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/sqrt(S+B)\",\"0.95\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/sqrt(S+B) @B_Purity=0.90", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/sqrt(S+B)\",\"0.90\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/sqrt(S+B) @B_Purity=0.85", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/sqrt(S+B)\",\"0.85\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/S", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/S\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/S @B_Purity=0.95", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/S\",\"0.95\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/S @B_Purity=0.90", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/S\",\"0.90\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/S @B_Purity=0.85", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/S\",\"0.85\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Before-Cut Efficiencies: B/S @B_Purity=0.80", ++ic ),
Form( ".x mvaeffs_beforecut.C+(\"%s\",\"B/S\",\"0.80\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ia) Classifier Cut Efficiencies: S/B ", ++ic ),
Form( ".x mvaeffs.C+(\"%s\",\"S/B\")", fName),
"Plots signal and background efficiencies versus cut on classifier output (macro mvaeffs.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%ib) Classifier Background Rejection vs Signal Efficiency (ROC curve)", ic ),
Form( ".x efficiencies.C(\"%s\")", fName ),
"Plots background rejection vs signal efficiencies (macro efficiencies.C) [\"ROC\" stands for \"Receiver Operation Characteristics\"]",
buttonType, defaultRequiredClassifier );
TString title = Form( "(%i) Parallel Coordinates (requires ROOT-version >= 5.17)", ++ic );
ActionButton( cbar,
title,
Form( ".x paracoor.C(\"%s\")", fName ),
"Plots parallel coordinates for classifiers and input variables (macro paracoor.C, requires ROOT >= 5.17)",
buttonType, defaultRequiredClassifier );
// parallel coordinates only exist since ROOT 5.17
#if ROOT_VERSION_CODE < ROOT_VERSION(5,17,0)
TMVAGui_inactiveButtons.push_back( title );
#endif
ActionButton( cbar,
Form( "(%i) PDFs of Classifiers (requires \"CreateMVAPdfs\" option set)", ++ic ),
Form( ".x probas.C(\"%s\")", fName ),
"Plots the PDFs of the classifier output distributions for signal and background - if requested (macro probas.C)",
buttonType, defaultRequiredClassifier );
ActionButton( cbar,
Form( "(%i) Likelihood Reference Distributiuons", ++ic),
Form( ".x likelihoodrefs.C(\"%s\")", fName ),
"Plots to verify the likelihood reference distributions (macro likelihoodrefs.C)",
buttonType, "Likelihood" );
ActionButton( cbar,
Form( "(%ia) Network Architecture (MLP)", ++ic ),
Form( ".x network.C(\"%s\")", fName ),
"Plots the MLP weights (macro network.C)",
buttonType, "MLP" );
ActionButton( cbar,
Form( "(%ib) Network Convergence Test (MLP)", ic ),
Form( ".x annconvergencetest.C(\"%s\")", fName ),
"Plots error estimator versus training epoch for training and test samples (macro annconvergencetest.C)",
buttonType, "MLP" );
ActionButton( cbar,
Form( "(%i) Decision Trees (BDT)", ++ic ),
Form( ".x BDT.C+(\"%s\")", fName ),
"Plots the Decision Trees trained by BDT algorithms (macro BDT.C(itree,...))",
buttonType, "BDT" );
ActionButton( cbar,
Form( "(%i) Decision Tree Control Plots (BDT)", ++ic ),
Form( ".x BDTControlPlots.C(\"%s\")", fName ),
"Plots to monitor boosting and pruning of decision trees (macro BDTControlPlots.C)",
buttonType, "BDT" );
// ActionButton( cbar,
// Form( "(%i) Monitor Decision Tree Boosting", ++ic ),
// Form( ".x BDTBoostWatch.C+(\"%s\")", fName ),
// "Plots separation obtained by BDT method as function of boosting steps (macro BDTBoostWatch.C(itree,...))",
// buttonType, "BDT" );
// ActionButton( cbar,
// Form( "(%i) Rule Ensemble Importance Plots (RuleFit)", ++ic ),
// Form( ".x rulevis.C(\"%s\",0)", fName ),
// "Plots all input variables with rule ensemble weights, including linear terms (macro rulevis.C)",
// buttonType, "RuleFit" );
ActionButton( cbar,
Form( "(%i) Plot Foams (PDEFoam)", ++ic ),
Form( ".x PlotFoams.C", fName ),
"Plot Foams (macro PlotFoams.C)",
buttonType, "PDEFoam" );
ActionButton( cbar,
Form( "(%i) General Boost Control Plots", ++ic ),
Form( ".x BoostControlPlots.C(\"%s\")", fName ),
"Plots to monitor boosting of general classifiers (macro BoostControlPlots.C)",
buttonType, "Boost" );
cbar->AddSeparator();
cbar->AddButton( Form( "(%i) Quit", ++ic ), ".q", "Quit", buttonType );
// set the style
cbar->SetTextColor("black");
// there seems to be a bug in ROOT: font jumps back to default after pressing on >2 different buttons
// cbar->SetFont("-adobe-helvetica-bold-r-*-*-12-*-*-*-*-*-iso8859-1");
// draw
cbar->Show();
// indicate inactive buttons
for (UInt_t i=0; i<TMVAGui_inactiveButtons.size(); i++) cbar->SetButtonState( TMVAGui_inactiveButtons[i], 3 );
if (TMVAGui_inactiveButtons.size() > 0) {
cout << "=== Note: inactive buttons indicate that the corresponding classifiers were not trained ===" << endl;
}
gROOT->SaveContext();
}
