// @(#)root/tmva $Id$   
// Author: Andreas Hoecker, Joerg Stelzer, Helge Voss

/**********************************************************************************
 * Project: TMVA - a Root-integrated toolkit for multivariate data analysis       *
 * Package: TMVA                                                                  *
 * Class  : SeparationBase                                                        *
 * Web    : http://tmva.sourceforge.net                                           *
 *                                                                                *
 * Description: An interface to different separation critiera useded in various   *
 *              training algorithms, as there are:                                *
 *                                                                                *
 *          There are two things: the Separation Index, and the Separation Gain   *
 *          Separation Index:                                                     *
 *          Measure of the "purity" of a sample. If all elements (events) in the  *
 *          sample belong to the same class (e.g. signal or backgr), than the     *
 *          separation index is 0 (meaning 100% purity (or 0% purity as it is     *
 *          symmetric. The index becomes maximal, for perfectly mixed samples     *
 *          eg. purity=50% , N_signal = N_bkg                                     *
 *                                                                                *
 *          Separation Gain:                                                      *
 *          the measure of how the quality of separation of the sample increases  *
 *          by splitting the sample e.g. into a "left-node" and a "right-node"    *
 *          (N * Index_parent) - (N_left * Index_left) - (N_right * Index_right)  *
 *          this is then the quality crition which is optimized for when trying   *
 *          to increase the information in the system (making the best selection  *
 *                                                                                *
 * Authors (alphabetical):                                                        *
 *      Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland              *
 *      Helge Voss      <Helge.Voss@cern.ch>     - MPI-K Heidelberg, Germany      *
 *      Kai Voss        <Kai.Voss@cern.ch>       - U. of Victoria, Canada         *
 *                                                                                *
 * Copyright (c) 2005:                                                            *
 *      CERN, Switzerland                                                         * 
 *      U. of Victoria, Canada                                                    * 
 *      Heidelberg U., Germany                                                    * 
 *                                                                                *
 * Redistribution and use in source and binary forms, with or without             *
 * modification, are permitted according to the terms listed in LICENSE           *
 * (http://ttmva.sourceforge.net/LICENSE)                                         *
 **********************************************************************************/

#include "TMVA/SeparationBase.h"

ClassImp(TMVA::SeparationBase)

#include <limits>
#include <iostream>
#include "TMath.h"


TMVA::SeparationBase::SeparationBase() :
   fName(""),
   fPrecisionCut(TMath::Sqrt(std::numeric_limits<double>::epsilon()))
{
   // default constructor
}

//copy constructor
TMVA::SeparationBase::SeparationBase( const SeparationBase& s ) :
   fName(s.fName),
   fPrecisionCut(TMath::Sqrt(std::numeric_limits<double>::epsilon()))
{
   // copy constructor
}

//_______________________________________________________________________
Double_t TMVA::SeparationBase::GetSeparationGain(const Double_t &nSelS, const Double_t& nSelB,
                                                 const Double_t& nTotS, const Double_t& nTotB)
{
   // Separation Gain:
   // the measure of how the quality of separation of the sample increases
   // by splitting the sample e.g. into a "left-node" and a "right-node"
   // (N * Index_parent) - (N_left * Index_left) - (N_right * Index_right)
   // this is then the quality crition which is optimized for when trying
   // to increase the information in the system (making the best selection

   if ( (nTotS-nSelS)==nSelS && (nTotB-nSelB)==nSelB) return 0.;

   // Double_t parentIndex = (nTotS+nTotB) *this->GetSeparationIndex(nTotS,nTotB);

   // Double_t leftIndex   = ( ((nTotS - nSelS) + (nTotB - nSelB))
   //                          * this->GetSeparationIndex(nTotS-nSelS,nTotB-nSelB) );
   // Double_t rightIndex  = (nSelS+nSelB) * this->GetSeparationIndex(nSelS,nSelB);


   Double_t parentIndex = this->GetSeparationIndex(nTotS,nTotB);

   Double_t leftIndex   = ( ((nTotS - nSelS) + (nTotB - nSelB))/(nTotS+nTotB)
                            * this->GetSeparationIndex(nTotS-nSelS,nTotB-nSelB) );
   Double_t rightIndex  = (nSelS+nSelB)/(nTotS+nTotB) * this->GetSeparationIndex(nSelS,nSelB);

   Double_t diff = parentIndex - leftIndex - rightIndex;
   //Double_t diff = (parentIndex - leftIndex - rightIndex)/(nTotS+nTotB);

   if(diff<fPrecisionCut ) {
      // std::cout << " Warning value in GetSeparation is below numerical presicion " 
      //           << diff/parentIndex 
      //           << std::endl;
      return 0;
   }

   return diff;
}


 SeparationBase.cxx:1
 SeparationBase.cxx:2
 SeparationBase.cxx:3
 SeparationBase.cxx:4
 SeparationBase.cxx:5
 SeparationBase.cxx:6
 SeparationBase.cxx:7
 SeparationBase.cxx:8
 SeparationBase.cxx:9
 SeparationBase.cxx:10
 SeparationBase.cxx:11
 SeparationBase.cxx:12
 SeparationBase.cxx:13
 SeparationBase.cxx:14
 SeparationBase.cxx:15
 SeparationBase.cxx:16
 SeparationBase.cxx:17
 SeparationBase.cxx:18
 SeparationBase.cxx:19
 SeparationBase.cxx:20
 SeparationBase.cxx:21
 SeparationBase.cxx:22
 SeparationBase.cxx:23
 SeparationBase.cxx:24
 SeparationBase.cxx:25
 SeparationBase.cxx:26
 SeparationBase.cxx:27
 SeparationBase.cxx:28
 SeparationBase.cxx:29
 SeparationBase.cxx:30
 SeparationBase.cxx:31
 SeparationBase.cxx:32
 SeparationBase.cxx:33
 SeparationBase.cxx:34
 SeparationBase.cxx:35
 SeparationBase.cxx:36
 SeparationBase.cxx:37
 SeparationBase.cxx:38
 SeparationBase.cxx:39
 SeparationBase.cxx:40
 SeparationBase.cxx:41
 SeparationBase.cxx:42
 SeparationBase.cxx:43
 SeparationBase.cxx:44
 SeparationBase.cxx:45
 SeparationBase.cxx:46
 SeparationBase.cxx:47
 SeparationBase.cxx:48
 SeparationBase.cxx:49
 SeparationBase.cxx:50
 SeparationBase.cxx:51
 SeparationBase.cxx:52
 SeparationBase.cxx:53
 SeparationBase.cxx:54
 SeparationBase.cxx:55
 SeparationBase.cxx:56
 SeparationBase.cxx:57
 SeparationBase.cxx:58
 SeparationBase.cxx:59
 SeparationBase.cxx:60
 SeparationBase.cxx:61
 SeparationBase.cxx:62
 SeparationBase.cxx:63
 SeparationBase.cxx:64
 SeparationBase.cxx:65
 SeparationBase.cxx:66
 SeparationBase.cxx:67
 SeparationBase.cxx:68
 SeparationBase.cxx:69
 SeparationBase.cxx:70
 SeparationBase.cxx:71
 SeparationBase.cxx:72
 SeparationBase.cxx:73
 SeparationBase.cxx:74
 SeparationBase.cxx:75
 SeparationBase.cxx:76
 SeparationBase.cxx:77
 SeparationBase.cxx:78
 SeparationBase.cxx:79
 SeparationBase.cxx:80
 SeparationBase.cxx:81
 SeparationBase.cxx:82
 SeparationBase.cxx:83
 SeparationBase.cxx:84
 SeparationBase.cxx:85
 SeparationBase.cxx:86
 SeparationBase.cxx:87
 SeparationBase.cxx:88
 SeparationBase.cxx:89
 SeparationBase.cxx:90
 SeparationBase.cxx:91
 SeparationBase.cxx:92
 SeparationBase.cxx:93
 SeparationBase.cxx:94
 SeparationBase.cxx:95
 SeparationBase.cxx:96
 SeparationBase.cxx:97
 SeparationBase.cxx:98
 SeparationBase.cxx:99
 SeparationBase.cxx:100
 SeparationBase.cxx:101
 SeparationBase.cxx:102
 SeparationBase.cxx:103
 SeparationBase.cxx:104
 SeparationBase.cxx:105
 SeparationBase.cxx:106