faif::ml::DecisionTree< Val > Class Template Reference

Decision Tree Classifier. More...

#include <DecisionTree.hpp>

Inheritance diagram for faif::ml::DecisionTree< Val >:

Collaboration diagram for faif::ml::DecisionTree< Val >:

Public Types
typedef Classifier< Val >::AttrValue	AttrValue
typedef Classifier< Val >::AttrDomain	AttrDomain
typedef Classifier< Val >::AttrIdd	AttrIdd
typedef Classifier< Val >::AttrIddSerialize	AttrIddSerialize
typedef Classifier< Val >::Domains	Domains
typedef Classifier< Val >::Beliefs	Beliefs
typedef Classifier< Val >::ExampleTest	ExampleTest
typedef Classifier< Val >::ExampleTrain	ExampleTrain
typedef Classifier< Val >::ExamplesTrain	ExamplesTrain
typedef Val	Value

Public Member Functions
	DecisionTree (const Domains &attr_domains, const AttrDomain &category_domain)
virtual void	reset ()
virtual void	train (const ExamplesTrain &e)
	learn classifier (on the collection of training examples). More...
virtual AttrIdd	getCategory (const ExampleTest &) const
virtual Beliefs	getCategories (const ExampleTest &) const
	classify and return all classes with belief that the example is from given class More...
virtual void	write (std::ostream &os) const
const DecisionTreeTrainParams &	getTrainParams () const
void	setTrainParams (const DecisionTreeTrainParams &p)
void	prune (const ExamplesTrain &e)
	prune tree - plase not use the example set used for training More...
const Domains &	getAttrDomains () const
	accessor More...
const AttrDomain &	getCategoryDomain () const
	accessor More...
AttrIdd	getCategoryIdd (const AttrValue &val) const
	accessor (helper) More...

Friends
class	boost::serialization::access
	serialization using boost::serialization More...

Detailed Description

template<typename Val>
class faif::ml::DecisionTree< Val >

Decision Tree Classifier.

Contains the attributes, attribute values and categories, train examples, test examples and classifier methods.

Member Function Documentation

template<typename Val >

void faif::ml::DecisionTree< Val >::reset ( )

virtual

clear the tree

Implements faif::ml::Classifier< Val >.

Referenced by faif::ml::RandomForest< Val >::setTrainParams().

template<typename Val >

void faif::ml::DecisionTree< Val >::train ( const ExamplesTrain &  e )

virtual

learn classifier (on the collection of training examples).

learn classifier (on the collection of training examples), the decision tree using given train examples

Parameters

e	training examples collection
ALLOWED_NBR_MISC_EX	allowed number of badly classified examples for each category

Implements faif::ml::Classifier< Val >.

Referenced by faif::ml::RandomForest< Val >::setTrainParams().

template<typename Val >

DecisionTree< Val >::AttrIdd faif::ml::DecisionTree< Val >::getCategory ( const ExampleTest &  e ) const

virtual

classify

classify - return the major category for best node from decision tree

Implements faif::ml::Classifier< Val >.

Referenced by faif::ml::RandomForest< Val >::setTrainParams().

template<typename Val >

DecisionTree< Val >::Beliefs faif::ml::DecisionTree< Val >::getCategories ( const ExampleTest &  e ) const

virtual

classify and return all classes with belief that the example is from given class

Implements faif::ml::Classifier< Val >.

Referenced by faif::ml::RandomForest< Val >::setTrainParams().

template<typename Val >

void faif::ml::DecisionTree< Val >::write ( std::ostream &  os ) const

virtual

the ostream method

ostream method

Reimplemented from faif::ml::Classifier< Val >.

Referenced by faif::ml::DecisionTree< Val >::prune().

template<typename Val>

const DecisionTreeTrainParams& faif::ml::DecisionTree< Val >::getTrainParams ( ) const

inline

accessor - get training parameters

template<typename Val>

void faif::ml::DecisionTree< Val >::setTrainParams ( const DecisionTreeTrainParams &  p )

inline

mutator - set training parameters

template<typename Val >

void faif::ml::DecisionTree< Val >::prune

(

const ExamplesTrain &

e

)

prune tree - plase not use the example set used for training

Return the (smart)pointer to node which replace the old one. If no prunning is performed the input pointer and the output are the same.

bottom-up method, the uneven distribution of categories is not considered

References faif::ml::calcEntropy(), faif::ml::TrainExampleCategoryCounters< Val >::entropy(), faif::ml::TrainExampleCategoryCounters< Val >::get(), faif::ml::TrainExampleCategoryCounters< Val >::getHistogram(), faif::ml::TrainExampleCategoryCounters< Val >::getSum(), and faif::ml::DecisionTree< Val >::write().

template<typename Val>

const Domains& faif::ml::Classifier< Val >::getAttrDomains ( ) const

inlineinherited

accessor

Referenced by faif::ml::createExample(), faif::ml::createExampleStrict(), and faif::ml::RandomForest< Val >::getCategories().

template<typename Val>

const AttrDomain& faif::ml::Classifier< Val >::getCategoryDomain ( ) const

inlineinherited

accessor

Referenced by faif::ml::createExample().

template<typename Val>

AttrIdd faif::ml::Classifier< Val >::getCategoryIdd ( const AttrValue &  val ) const

inlineinherited

accessor (helper)

References faif::ml::Classifier< Val >::getCategories(), faif::ml::Classifier< Val >::getCategory(), faif::ml::Classifier< Val >::reset(), faif::ml::Classifier< Val >::train(), and faif::ml::Classifier< Val >::write().

Friends And Related Function Documentation

template<typename Val>

friend class boost::serialization::access

friend

serialization using boost::serialization

---

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

• faif/learning/DecisionTree.hpp