ROOT   Reference Guide
ExpectedErrorPruneTool.h
Go to the documentation of this file.
1/**********************************************************************************
2 * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
3 * Package: TMVA *
4 * Class : TMVA::DecisionTree *
5 * Web : http://tmva.sourceforge.net *
6 * *
7 * Description: *
8 * Implementation of a Decision Tree *
9 * *
10 * Authors (alphabetical): *
11 * Andreas Hoecker <Andreas.Hocker@cern.ch> - CERN, Switzerland *
12 * Helge Voss <Helge.Voss@cern.ch> - MPI-K Heidelberg, Germany *
13 * Kai Voss <Kai.Voss@cern.ch> - U. of Victoria, Canada *
14 * Doug Schouten <dschoute@sfu.ca> - Simon Fraser U., Canada *
15 * *
16 * Copyright (c) 2005: *
17 * CERN, Switzerland *
18 * U. of Victoria, Canada *
19 * MPI-K Heidelberg, Germany *
20 * *
21 * Redistribution and use in source and binary forms, with or without *
22 * modification, are permitted according to the terms listed in LICENSE *
24 * *
25 **********************************************************************************/
26
27#ifndef ROOT_TMVA_ExpectedErrorPruneTool
28#define ROOT_TMVA_ExpectedErrorPruneTool
29
30/////////////////////////////////////////////////////////////////////////////////////////////////////////////
31// ExpectedErrorPruneTool - a helper class to prune a decision tree using the expected error (C4.5) method //
32// //
33// Uses an upper limit on the error made by the classification done by each node. If the S/S+B of the node //
34// is f, then according to the training sample, the error rate (fraction of misclassified events by this //
35// node) is (1-f). Now f has a statistical error according to the binomial distribution hence the error on //
36// f can be estimated (same error as the binomial error for efficiency calculations //
37// ( sigma = sqrt(eff(1-eff)/nEvts ) ) //
38// //
39// This tool prunes branches from a tree if the expected error of a node is less than that of the sum of //
40// the error in its descendants. //
41// //
42/////////////////////////////////////////////////////////////////////////////////////////////////////////////
43
44#include <vector>
45
46#include "TMath.h"
47
48#include "TMVA/IPruneTool.h"
49
50namespace TMVA {
51
52 class MsgLogger;
53
55 public:
57 virtual ~ExpectedErrorPruneTool( );
58
59 // returns the PruningInfo object for a given tree and test sample
60 virtual PruningInfo* CalculatePruningInfo( DecisionTree* dt, const IPruneTool::EventSample* testEvents = NULL,
61 Bool_t isAutomatic = kFALSE );
62
63 public:
64 // set the increment dalpha with which to scan for the optimal prune strength
65 inline void SetPruneStrengthIncrement( Double_t dalpha ) { fDeltaPruneStrength = dalpha; }
66
67 private:
71 Int_t CountNodes( DecisionTreeNode* node, Int_t icount = 0 );
72
73 Double_t fDeltaPruneStrength; //! the stepsize for optimizing the pruning strength parameter
74 Double_t fNodePurityLimit; //! the purity limit for labelling a terminal node as signal
75 std::vector<DecisionTreeNode*> fPruneSequence; //! the (optimal) prune sequence
76 // std::multimap<const Double_t, Double_t> fQualityMap; //! map of tree quality <=> prune strength
77 mutable MsgLogger* fLogger; // message logger
78 MsgLogger& Log() const { return *fLogger; }
79 };
80
84 Int_t counter = icount + 1; // count this node
85 if(!(node->IsTerminal()) && l != NULL && r != NULL) {
86 counter = CountNodes(l,counter);
87 counter = CountNodes(r,counter);
88 }
89 return counter;
90 }
91}
92
93#endif
94
ROOT::R::TRInterface & r
Definition: Object.C:4
const Bool_t kFALSE
Definition: RtypesCore.h:90
double Double_t
Definition: RtypesCore.h:57
virtual DecisionTreeNode * GetLeft() const
virtual DecisionTreeNode * GetRight() const
Implementation of a Decision Tree.
Definition: DecisionTree.h:64
A helper class to prune a decision tree using the expected error (C4.5) method.
Double_t fNodePurityLimit
the stepsize for optimizing the pruning strength parameter
void FindListOfNodes(DecisionTreeNode *node)
recursive pruning of nodes using the Expected Error Pruning (EEP)
Double_t GetNodeError(DecisionTreeNode *node) const
Calculate an UPPER limit on the error made by the classification done by this node.
Double_t GetSubTreeError(DecisionTreeNode *node) const
calculate the expected statistical error on the subtree below "node" which is used in the expected er...
virtual PruningInfo * CalculatePruningInfo(DecisionTree *dt, const IPruneTool::EventSample *testEvents=NULL, Bool_t isAutomatic=kFALSE)
void SetPruneStrengthIncrement(Double_t dalpha)
std::vector< DecisionTreeNode * > fPruneSequence
the purity limit for labelling a terminal node as signal
Int_t CountNodes(DecisionTreeNode *node, Int_t icount=0)
MsgLogger * fLogger
the (optimal) prune sequence
IPruneTool - a helper interface class to prune a decision tree.
Definition: IPruneTool.h:70
std::vector< const Event * > EventSample
Definition: IPruneTool.h:74
ostringstream derivative to redirect and format output
Definition: MsgLogger.h:59
create variable transformations
auto * l
Definition: textangle.C:4