TMVA::CCPruner Class Reference

A helper class to prune a decision tree using the Cost Complexity method (see Classification and Regression Trees by Leo Breiman et al)

Some definitions:

• \( T_{max} \) - the initial, usually highly overtrained tree, that is to be pruned back
• \( R(T) \) - quality index (Gini, misclassification rate, or other) of a tree \( T \)
• \( \sim T \) - set of terminal nodes in \( T \)
• \( T' \) - the pruned subtree of \( T_max \) that has the best quality index \( R(T') \)
• \( \alpha \) - the prune strength parameter in Cost Complexity pruning \( (R_{\alpha}(T) = R(T) + \alpha*|\sim T|) \)

There are two running modes in CCPruner: (i) one may select a prune strength and prune back the tree \( T_{max}\) until the criterion:

\[ \alpha < \frac{R(T) - R(t)}{|\sim T_t| - 1} \]

is true for all nodes t in \( T \), or (ii) the algorithm finds the sequence of critical points \( \alpha_k < \alpha_{k+1} ... < \alpha_K \) such that \( T_K = root(T_{max}) \) and then selects the optimally-pruned subtree, defined to be the subtree with the best quality index for the validation sample.

Definition at line 61 of file CCPruner.h.

Public Types
typedef std::vector< Event * >	EventList

Public Member Functions
	CCPruner (DecisionTree *t_max, const EventList *validationSample, SeparationBase *qualityIndex=NULL)
	constructor More...
	CCPruner (DecisionTree *t_max, const DataSet *validationSample, SeparationBase *qualityIndex=NULL)
	constructor More...
	~CCPruner ()
std::vector< TMVA::DecisionTreeNode * >	GetOptimalPruneSequence () const
	return the prune strength (=alpha) corresponding to the prune sequence More...
Float_t	GetOptimalPruneStrength () const
Float_t	GetOptimalQualityIndex () const
void	Optimize ()
	determine the pruning sequence More...
void	SetPruneStrength (Float_t alpha=-1.0)

Private Attributes
Float_t	fAlpha
Bool_t	fDebug
	index of the optimal tree in the pruned tree sequence More...
Int_t	fOptimalK
	map of R(T) -> pruning index More...
Bool_t	fOwnQIndex
	the quality index used to calculate R(t), R(T) = sum[t in ~T]{ R(t) } More...
std::vector< TMVA::DecisionTreeNode * >	fPruneSequence
	(pruned) decision tree More...
std::vector< Float_t >	fPruneStrengthList
	map of weakest links (i.e., branches to prune) -> pruning index More...
SeparationBase *	fQualityIndex
	the event sample to select the optimally-pruned tree More...
std::vector< Float_t >	fQualityIndexList
	map of alpha -> pruning index More...
DecisionTree *	fTree
	flag indicates if fQualityIndex is owned by this More...
const DataSet *	fValidationDataSet
	the event sample to select the optimally-pruned tree More...
const EventList *	fValidationSample
	regularization parameter in CC pruning More...

#include <TMVA/CCPruner.h>

Member Typedef Documentation

EventList

typedef std::vector<Event*> TMVA::CCPruner::EventList

Definition at line 63 of file CCPruner.h.

Constructor & Destructor Documentation

CCPruner() [1/2]

CCPruner::CCPruner	(	DecisionTree *	t_max,
		const EventList *	validationSample,
		SeparationBase *	qualityIndex = NULL
	)

constructor

Definition at line 69 of file CCPruner.cxx.

CCPruner() [2/2]

CCPruner::CCPruner	(	DecisionTree *	t_max,
		const DataSet *	validationSample,
		SeparationBase *	qualityIndex = NULL
	)

constructor

Definition at line 92 of file CCPruner.cxx.

~CCPruner()

CCPruner::~CCPruner

(

)

Definition at line 115 of file CCPruner.cxx.

Member Function Documentation

GetOptimalPruneSequence()

std::vector< DecisionTreeNode * > CCPruner::GetOptimalPruneSequence

(

)

const

return the prune strength (=alpha) corresponding to the prune sequence

Definition at line 240 of file CCPruner.cxx.

GetOptimalPruneStrength()

Float_t TMVA::CCPruner::GetOptimalPruneStrength ( ) const

inline

Definition at line 88 of file CCPruner.h.

GetOptimalQualityIndex()

Float_t TMVA::CCPruner::GetOptimalQualityIndex ( ) const

inline

Definition at line 84 of file CCPruner.h.

Optimize()

void CCPruner::Optimize

(

)

determine the pruning sequence

Definition at line 124 of file CCPruner.cxx.

SetPruneStrength()

void TMVA::CCPruner::SetPruneStrength ( Float_t  alpha = -1.0 )

inline

Definition at line 109 of file CCPruner.h.

Member Data Documentation

fAlpha

Float_t TMVA::CCPruner::fAlpha

private

Definition at line 92 of file CCPruner.h.

fDebug

Bool_t TMVA::CCPruner::fDebug

private

index of the optimal tree in the pruned tree sequence

Definition at line 105 of file CCPruner.h.

fOptimalK

Int_t TMVA::CCPruner::fOptimalK

private

map of R(T) -> pruning index

Definition at line 104 of file CCPruner.h.

fOwnQIndex

Bool_t TMVA::CCPruner::fOwnQIndex

private

the quality index used to calculate R(t), R(T) = sum[t in ~T]{ R(t) }

Definition at line 96 of file CCPruner.h.

fPruneSequence

std::vector<TMVA::DecisionTreeNode*> TMVA::CCPruner::fPruneSequence

private

(pruned) decision tree

Definition at line 100 of file CCPruner.h.

fPruneStrengthList

std::vector<Float_t> TMVA::CCPruner::fPruneStrengthList

private

map of weakest links (i.e., branches to prune) -> pruning index

Definition at line 101 of file CCPruner.h.

fQualityIndex

SeparationBase* TMVA::CCPruner::fQualityIndex

private

the event sample to select the optimally-pruned tree

Definition at line 95 of file CCPruner.h.

fQualityIndexList

std::vector<Float_t> TMVA::CCPruner::fQualityIndexList

private

map of alpha -> pruning index

Definition at line 102 of file CCPruner.h.

fTree

DecisionTree* TMVA::CCPruner::fTree

private

flag indicates if fQualityIndex is owned by this

Definition at line 98 of file CCPruner.h.

fValidationDataSet

const DataSet* TMVA::CCPruner::fValidationDataSet

private

the event sample to select the optimally-pruned tree

Definition at line 94 of file CCPruner.h.

fValidationSample

const EventList* TMVA::CCPruner::fValidationSample

private

regularization parameter in CC pruning

Definition at line 93 of file CCPruner.h.

---

The documentation for this class was generated from the following files:

• tmva/tmva/inc/TMVA/CCPruner.h
• tmva/tmva/src/CCPruner.cxx