| 1 |
/**********************************************************************************
|
| 2 |
* Project : TMVA - a Root-integrated toolkit for multivariate data analysis *
|
| 3 |
* Package : TMVA *
|
| 4 |
* Exectuable: TMVARegressionApplication *
|
| 5 |
* *
|
| 6 |
* This macro provides a simple example on how to use the trained regression MVAs *
|
| 7 |
* within an analysis module *
|
| 8 |
**********************************************************************************/
|
| 9 |
|
| 10 |
#include <cstdlib>
|
| 11 |
#include <vector>
|
| 12 |
#include <iostream>
|
| 13 |
#include <map>
|
| 14 |
#include <string>
|
| 15 |
|
| 16 |
#include "TFile.h"
|
| 17 |
#include "TTree.h"
|
| 18 |
#include "TString.h"
|
| 19 |
#include "TSystem.h"
|
| 20 |
#include "TROOT.h"
|
| 21 |
#include "TStopwatch.h"
|
| 22 |
|
| 23 |
#if not defined(__CINT__) || defined(__MAKECINT__)
|
| 24 |
#include "TMVA/Tools.h"
|
| 25 |
#include "TMVA/Reader.h"
|
| 26 |
#endif
|
| 27 |
|
| 28 |
using namespace TMVA;
|
| 29 |
|
| 30 |
void TMVARegressionApplication( TString myMethodList = "" )
|
| 31 |
{
|
| 32 |
//---------------------------------------------------------------
|
| 33 |
// This loads the library
|
| 34 |
TMVA::Tools::Instance();
|
| 35 |
|
| 36 |
// Default MVA methods to be trained + tested
|
| 37 |
std::map<std::string,int> Use;
|
| 38 |
|
| 39 |
// --- Mutidimensional likelihood and Nearest-Neighbour methods
|
| 40 |
Use["PDERS"] = 0;
|
| 41 |
Use["PDEFoam"] = 1;
|
| 42 |
Use["KNN"] = 1;
|
| 43 |
//
|
| 44 |
// --- Linear Discriminant Analysis
|
| 45 |
Use["LD"] = 1;
|
| 46 |
//
|
| 47 |
// --- Function Discriminant analysis
|
| 48 |
Use["FDA_GA"] = 1;
|
| 49 |
Use["FDA_MC"] = 0;
|
| 50 |
Use["FDA_MT"] = 0;
|
| 51 |
Use["FDA_GAMT"] = 0;
|
| 52 |
//
|
| 53 |
// --- Neural Network
|
| 54 |
Use["MLP"] = 1;
|
| 55 |
//
|
| 56 |
// --- Support Vector Machine
|
| 57 |
Use["SVM"] = 0;
|
| 58 |
//
|
| 59 |
// --- Boosted Decision Trees
|
| 60 |
Use["BDT"] = 0;
|
| 61 |
Use["BDTG"] = 1;
|
| 62 |
// ---------------------------------------------------------------
|
| 63 |
|
| 64 |
std::cout << std::endl;
|
| 65 |
std::cout << "==> Start TMVARegressionApplication" << std::endl;
|
| 66 |
|
| 67 |
// Select methods (don't look at this code - not of interest)
|
| 68 |
if (myMethodList != "") {
|
| 69 |
for (std::map<std::string,int>::iterator it = Use.begin(); it != Use.end(); it++) it->second = 0;
|
| 70 |
|
| 71 |
std::vector<TString> mlist = gTools().SplitString( myMethodList, ',' );
|
| 72 |
for (UInt_t i=0; i<mlist.size(); i++) {
|
| 73 |
std::string regMethod(mlist[i]);
|
| 74 |
|
| 75 |
if (Use.find(regMethod) == Use.end()) {
|
| 76 |
std::cout << "Method \"" << regMethod << "\" not known in TMVA under this name. Choose among the following:" << std::endl;
|
| 77 |
for (std::map<std::string,int>::iterator it = Use.begin(); it != Use.end(); it++) std::cout << it->first << " ";
|
| 78 |
std::cout << std::endl;
|
| 79 |
return;
|
| 80 |
}
|
| 81 |
Use[regMethod] = 1;
|
| 82 |
}
|
| 83 |
}
|
| 84 |
|
| 85 |
// --------------------------------------------------------------------------------------------------
|
| 86 |
|
| 87 |
// --- Create the Reader object
|
| 88 |
|
| 89 |
TMVA::Reader *reader = new TMVA::Reader( "!Color:!Silent" );
|
| 90 |
|
| 91 |
// Create a set of variables and declare them to the reader
|
| 92 |
// - the variable names MUST corresponds in name and type to those given in the weight file(s) used
|
| 93 |
Float_t var1, var2;
|
| 94 |
reader->AddVariable( "var1", &var1 );
|
| 95 |
reader->AddVariable( "var2", &var2 );
|
| 96 |
|
| 97 |
// Spectator variables declared in the training have to be added to the reader, too
|
| 98 |
Float_t spec1,spec2;
|
| 99 |
reader->AddSpectator( "spec1:=var1*2", &spec1 );
|
| 100 |
reader->AddSpectator( "spec2:=var1*3", &spec2 );
|
| 101 |
|
| 102 |
// --- Book the MVA methods
|
| 103 |
|
| 104 |
TString dir = "weights/";
|
| 105 |
TString prefix = "TMVARegression";
|
| 106 |
|
| 107 |
// Book method(s)
|
| 108 |
for (std::map<std::string,int>::iterator it = Use.begin(); it != Use.end(); it++) {
|
| 109 |
if (it->second) {
|
| 110 |
TString methodName = it->first + " method";
|
| 111 |
TString weightfile = dir + prefix + "_" + TString(it->first) + ".weights.xml";
|
| 112 |
reader->BookMVA( methodName, weightfile );
|
| 113 |
}
|
| 114 |
}
|
| 115 |
|
| 116 |
// Book output histograms
|
| 117 |
TH1* hists[100];
|
| 118 |
Int_t nhists = -1;
|
| 119 |
for (std::map<std::string,int>::iterator it = Use.begin(); it != Use.end(); it++) {
|
| 120 |
TH1* h = new TH1F( it->first.c_str(), TString(it->first) + " method", 100, -100, 600 );
|
| 121 |
if (it->second) hists[++nhists] = h;
|
| 122 |
}
|
| 123 |
nhists++;
|
| 124 |
|
| 125 |
// Prepare input tree (this must be replaced by your data source)
|
| 126 |
// in this example, there is a toy tree with signal and one with background events
|
| 127 |
// we'll later on use only the "signal" events for the test in this example.
|
| 128 |
//
|
| 129 |
TFile *input(0);
|
| 130 |
TString fname = "./tmva_reg_example.root";
|
| 131 |
if (!gSystem->AccessPathName( fname )) {
|
| 132 |
input = TFile::Open( fname ); // check if file in local directory exists
|
| 133 |
}
|
| 134 |
else {
|
| 135 |
input = TFile::Open( "http://root.cern.ch/files/tmva_reg_example.root" ); // if not: download from ROOT server
|
| 136 |
}
|
| 137 |
|
| 138 |
if (!input) {
|
| 139 |
std::cout << "ERROR: could not open data file" << std::endl;
|
| 140 |
exit(1);
|
| 141 |
}
|
| 142 |
std::cout << "--- TMVARegressionApp : Using input file: " << input->GetName() << std::endl;
|
| 143 |
|
| 144 |
// --- Event loop
|
| 145 |
|
| 146 |
// Prepare the tree
|
| 147 |
// - here the variable names have to corresponds to your tree
|
| 148 |
// - you can use the same variables as above which is slightly faster,
|
| 149 |
// but of course you can use different ones and copy the values inside the event loop
|
| 150 |
//
|
| 151 |
TTree* theTree = (TTree*)input->Get("TreeR");
|
| 152 |
std::cout << "--- Select signal sample" << std::endl;
|
| 153 |
theTree->SetBranchAddress( "var1", &var1 );
|
| 154 |
theTree->SetBranchAddress( "var2", &var2 );
|
| 155 |
|
| 156 |
std::cout << "--- Processing: " << theTree->GetEntries() << " events" << std::endl;
|
| 157 |
TStopwatch sw;
|
| 158 |
sw.Start();
|
| 159 |
for (Long64_t ievt=0; ievt<theTree->GetEntries();ievt++) {
|
| 160 |
|
| 161 |
if (ievt%1000 == 0) {
|
| 162 |
std::cout << "--- ... Processing event: " << ievt << std::endl;
|
| 163 |
}
|
| 164 |
|
| 165 |
theTree->GetEntry(ievt);
|
| 166 |
|
| 167 |
// Retrieve the MVA target values (regression outputs) and fill into histograms
|
| 168 |
// NOTE: EvaluateRegression(..) returns a vector for multi-target regression
|
| 169 |
|
| 170 |
for (Int_t ih=0; ih<nhists; ih++) {
|
| 171 |
TString title = hists[ih]->GetTitle();
|
| 172 |
Float_t val = (reader->EvaluateRegression( title ))[0];
|
| 173 |
hists[ih]->Fill( val );
|
| 174 |
}
|
| 175 |
}
|
| 176 |
sw.Stop();
|
| 177 |
std::cout << "--- End of event loop: "; sw.Print();
|
| 178 |
|
| 179 |
// --- Write histograms
|
| 180 |
|
| 181 |
TFile *target = new TFile( "TMVARegApp.root","RECREATE" );
|
| 182 |
for (Int_t ih=0; ih<nhists; ih++) hists[ih]->Write();
|
| 183 |
target->Close();
|
| 184 |
|
| 185 |
std::cout << "--- Created root file: \"" << target->GetName()
|
| 186 |
<< "\" containing the MVA output histograms" << std::endl;
|
| 187 |
|
| 188 |
delete reader;
|
| 189 |
|
| 190 |
std::cout << "==> TMVARegressionApplication is done!" << std::endl << std::endl;
|
| 191 |
}
|
