The HoeffdingTree object represents all of the necessary information for a Hoeffding-bound-based decision tree. More...

Public Types
typedef CategoricalSplitType< FitnessFunction >	CategoricalSplit
	Allow access to the categorical split type. More...

typedef NumericSplitType< FitnessFunction >	NumericSplit
	Allow access to the numeric split type. More...

Public Member Functions
template < typename MatType >
	HoeffdingTree (const MatType &data, const data::DatasetInfo &datasetInfo, const arma::Row< size_t > &labels, const size_t numClasses, const bool batchTraining=true, const double successProbability=0.95, const size_t maxSamples=0, const size_t checkInterval=100, const size_t minSamples=100, const CategoricalSplitType< FitnessFunction > &categoricalSplitIn=CategoricalSplitType< FitnessFunction >(0, 0), const NumericSplitType< FitnessFunction > &numericSplitIn=NumericSplitType< FitnessFunction >(0))
	Construct the Hoeffding tree with the given parameters and given training data. More...

	HoeffdingTree (const data::DatasetInfo &datasetInfo, const size_t numClasses, const double successProbability=0.95, const size_t maxSamples=0, const size_t checkInterval=100, const size_t minSamples=100, const CategoricalSplitType< FitnessFunction > &categoricalSplitIn=CategoricalSplitType< FitnessFunction >(0, 0), const NumericSplitType< FitnessFunction > &numericSplitIn=NumericSplitType< FitnessFunction >(0), std::unordered_map< size_t, std::pair< size_t, size_t >> *dimensionMappings=NULL, const bool copyDatasetInfo=true)
	Construct the Hoeffding tree with the given parameters, but training on no data. More...

	HoeffdingTree ()
	Construct a Hoeffding tree with no data and no information. More...

	HoeffdingTree (const HoeffdingTree &other)
	Copy another tree (warning: this will duplicate the tree entirely, and may use a lot of memory. More...

	HoeffdingTree (HoeffdingTree &&other)
	Move another tree. More...

	~HoeffdingTree ()
	Clean up memory. More...

template < typename VecType >
size_t	CalculateDirection (const VecType &point) const
	Given a point and that this node is not a leaf, calculate the index of the child node this point would go towards. More...

size_t	CheckInterval () const
	Get the number of samples before a split check is performed. More...

void	CheckInterval (const size_t checkInterval)
	Modify the number of samples before a split check is performed. More...

const HoeffdingTree &	Child (const size_t i) const
	Get a child. More...

HoeffdingTree &	Child (const size_t i)
	Modify a child. More...

template < typename VecType >
size_t	Classify (const VecType &point) const
	Classify the given point, using this node and the entire (sub)tree beneath it. More...

template < typename VecType >
void	Classify (const VecType &point, size_t &prediction, double &probability) const
	Classify the given point and also return an estimate of the probability that the prediction is correct. More...

template < typename MatType >
void	Classify (const MatType &data, arma::Row< size_t > &predictions) const
	Classify the given points, using this node and the entire (sub)tree beneath it. More...

template < typename MatType >
void	Classify (const MatType &data, arma::Row< size_t > &predictions, arma::rowvec &probabilities) const
	Classify the given points, using this node and the entire (sub)tree beneath it. More...

void	CreateChildren ()
	Given that this node should split, create the children. More...

size_t	MajorityClass () const
	Get the majority class. More...

size_t &	MajorityClass ()
	Modify the majority class. More...

double	MajorityProbability () const
	Get the probability of the majority class (based on training samples). More...

double &	MajorityProbability ()
	Modify the probability of the majority class. More...

size_t	MaxSamples () const
	Get the maximum number of samples before a split is forced. More...

void	MaxSamples (const size_t maxSamples)
	Modify the maximum number of samples before a split is forced. More...

size_t	MinSamples () const
	Get the minimum number of samples for a split. More...

void	MinSamples (const size_t minSamples)
	Modify the minimum number of samples for a split. More...

size_t	NumChildren () const
	Get the number of children. More...

size_t	NumDescendants () const
	Get the size of the Hoeffding Tree. More...

HoeffdingTree &	operator= (const HoeffdingTree &other)
	Copy assignment operator. More...

HoeffdingTree &	operator= (HoeffdingTree &&other)
	Move assignment operator. More...

template < typename Archive >
void	serialize (Archive &ar, const uint32_t)
	Serialize the split. More...

size_t	SplitCheck ()
	Check if a split would satisfy the conditions of the Hoeffding bound with the node's specified success probability. More...

size_t	SplitDimension () const
	Get the splitting dimension (size_t(-1) if no split). More...

double	SuccessProbability () const
	Get the confidence required for a split. More...

void	SuccessProbability (const double successProbability)
	Modify the confidence required for a split. More...

template < typename MatType >
void	Train (const MatType &data, const arma::Row< size_t > &labels, const bool batchTraining=true, const bool resetTree=false, const size_t numClasses=0)
	Train on a set of points, either in streaming mode or in batch mode, with the given labels. More...

template < typename MatType >
void	Train (const MatType &data, const data::DatasetInfo &info, const arma::Row< size_t > &labels, const bool batchTraining=true, const size_t numClasses=0)
	Train on a set of points, either in streaming mode or in batch mode, with the given labels and the given `DatasetInfo`. More...

template < typename VecType >
void	Train (const VecType &point, const size_t label)
	Train on a single point in streaming mode, with the given label. More...

Detailed Description

template
<
typename
FitnessFunction
=
GiniImpurity
,
template
<
typename
>
class
NumericSplitType
=
HoeffdingDoubleNumericSplit
,
template
<
typename
>
class
CategoricalSplitType
=
HoeffdingCategoricalSplit
>

class mlpack::tree::HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >

The HoeffdingTree object represents all of the necessary information for a Hoeffding-bound-based decision tree.

This class is able to train on samples in streaming settings and batch settings, and perform splits based on the Hoeffding bound. The Hoeffding tree (also known as the "very fast decision tree" – VFDT) is described in the following paper:

@inproceedings{domingos2000mining,
    title={{Mining High-Speed Data Streams}},
    author={Domingos, P. and Hulten, G.},
    year={2000},
    booktitle={Proceedings of the Sixth ACM SIGKDD International Conference
        on Knowledge Discovery and Data Mining (KDD '00)},
    pages={71--80}
}

The class is modular, and takes three template parameters. The first, FitnessFunction, is the fitness function that should be used to determine whether a split is beneficial; examples might be GiniImpurity or HoeffdingInformationGain. The NumericSplitType determines how numeric attributes are handled, and the CategoricalSplitType determines how categorical attributes are handled. As far as the actual splitting goes, the meat of the splitting procedure will be contained in those two classes.

Template Parameters

FitnessFunction	Fitness function to use.
NumericSplitType	Technique for splitting numeric features.
CategoricalSplitType	Technique for splitting categorical features.

Definition at line 61 of file hoeffding_tree.hpp.

Member Typedef Documentation

◆ CategoricalSplit

typedef CategoricalSplitType<FitnessFunction> CategoricalSplit

Allow access to the categorical split type.

Definition at line 67 of file hoeffding_tree.hpp.

◆ NumericSplit

typedef NumericSplitType<FitnessFunction> NumericSplit

Allow access to the numeric split type.

Definition at line 65 of file hoeffding_tree.hpp.

Constructor & Destructor Documentation

◆ HoeffdingTree() [1/5]

HoeffdingTree	(	const MatType &	data,
		const data::DatasetInfo &	datasetInfo,
		const arma::Row< size_t > &	labels,
		const size_t	numClasses,
		const bool	batchTraining = `true`,
		const double	successProbability = `0.95`,
		const size_t	maxSamples = `0`,
		const size_t	checkInterval = `100`,
		const size_t	minSamples = `100`,
		const CategoricalSplitType< FitnessFunction > &	categoricalSplitIn = `CategoricalSplitType< FitnessFunction >(0, 0)`,
		const NumericSplitType< FitnessFunction > &	numericSplitIn = `NumericSplitType< FitnessFunction >(0)`
	)

Construct the Hoeffding tree with the given parameters and given training data.

The tree may be trained either in batch mode (which looks at all points before splitting, and propagates these points to the created children for further training), or in streaming mode, where each point is only considered once. (In general, batch mode will give better-performing trees, but will have higher memory and runtime costs for the same dataset.)

Parameters

data	Dataset to train on.
datasetInfo	Information on the dataset (types of each feature).
labels	Labels of each point in the dataset.
numClasses	Number of classes in the dataset.
batchTraining	Whether or not to train in batch.
successProbability	Probability of success required in Hoeffding bounds before a split can happen.
maxSamples	Maximum number of samples before a split is forced (0 never forces a split); ignored in batch training mode.
checkInterval	Number of samples required before each split; ignored in batch training mode.
minSamples	If the node has seen this many points or fewer, no split will be allowed.
categoricalSplitIn	Optional instantiated categorical split object.
numericSplitIn	Optional instantiated numeric split object.

◆ HoeffdingTree() [2/5]

HoeffdingTree	(	const data::DatasetInfo &	datasetInfo,
		const size_t	numClasses,
		const double	successProbability = `0.95`,
		const size_t	maxSamples = `0`,
		const size_t	checkInterval = `100`,
		const size_t	minSamples = `100`,
		const CategoricalSplitType< FitnessFunction > &	categoricalSplitIn = `CategoricalSplitType< FitnessFunction >(0, 0)`,
		const NumericSplitType< FitnessFunction > &	numericSplitIn = `NumericSplitType< FitnessFunction >(0)`,
		std::unordered_map< size_t, std::pair< size_t, size_t >> *	dimensionMappings = `NULL`,
		const bool	copyDatasetInfo = `true`
	)

Construct the Hoeffding tree with the given parameters, but training on no data.

The dimensionMappings parameter is only used if it is desired that this node does not create its own dimensionMappings object (for instance, if this is a child of another node in the tree).

Parameters

numClasses	Number of classes in the dataset.
datasetInfo	Information on the dataset (types of each feature).
successProbability	Probability of success required in Hoeffding bound before a split can happen.
maxSamples	Maximum number of samples before a split is forced.
checkInterval	Number of samples required before each split check.
minSamples	If the node has seen this many points or fewer, no split will be allowed.
dimensionMappings	Mappings from dimension indices to positions in numeric and categorical split vectors. If left NULL, a new one will be created.
copyDatasetInfo	If true, then a copy of the datasetInfo will be made.
categoricalSplitIn	Optional instantiated categorical split object.
numericSplitIn	Optional instantiated numeric split object.

◆ HoeffdingTree() [3/5]

HoeffdingTree ( )

Construct a Hoeffding tree with no data and no information.

Be sure to call Train() before trying to use the tree.

◆ HoeffdingTree() [4/5]

HoeffdingTree ( const HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType > & other )

Copy another tree (warning: this will duplicate the tree entirely, and may use a lot of memory.

Make sure it's what you want before you do it).

Parameters

other Tree to copy.

◆ HoeffdingTree() [5/5]

HoeffdingTree ( HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType > && other )

Move another tree.

Parameters

other Tree to move.

◆ ~HoeffdingTree()

~HoeffdingTree ( )

Clean up memory.

Member Function Documentation

◆ CalculateDirection()

size_t CalculateDirection ( const VecType & point ) const

Given a point and that this node is not a leaf, calculate the index of the child node this point would go towards.

This method is primarily used by the Classify() function, but it can be used in a standalone sense too.

Parameters

point Point to classify.

Referenced by HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CheckInterval().

◆ CheckInterval() [1/2]

size_t CheckInterval ( ) const

inline

Get the number of samples before a split check is performed.

Definition at line 284 of file hoeffding_tree.hpp.

References HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CalculateDirection(), HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::Classify(), HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CreateChildren(), HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::NumDescendants(), and HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::serialize().

◆ CheckInterval() [2/2]

void CheckInterval ( const size_t checkInterval )

Modify the number of samples before a split check is performed.

◆ Child() [1/2]

const HoeffdingTree& Child ( const size_t i ) const

inline

Get a child.

Definition at line 264 of file hoeffding_tree.hpp.

◆ Child() [2/2]

HoeffdingTree& Child ( const size_t i )

inline

Modify a child.

Definition at line 266 of file hoeffding_tree.hpp.

◆ Classify() [1/4]

size_t Classify ( const VecType & point ) const

Classify the given point, using this node and the entire (sub)tree beneath it.

The predicted label is returned.

Parameters

point Point to classify.

Returns: Predicted label of point.

Referenced by HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CheckInterval().

◆ Classify() [2/4]

void Classify	(	const VecType &	point,
		size_t &	prediction,
		double &	probability
	)		const

Classify the given point and also return an estimate of the probability that the prediction is correct.

(This estimate is simply the probability that a training point was from the majority class in the leaf that this point binned to.)

Parameters

point	Point to classify.
prediction	Predicted label of point.
probability	An estimate of the probability that the prediction is correct.

◆ Classify() [3/4]

void Classify	(	const MatType &	data,
		arma::Row< size_t > &	predictions
	)		const

Classify the given points, using this node and the entire (sub)tree beneath it.

The predicted labels for each point are returned.

Parameters

data	Points to classify.
predictions	Predicted labels for each point.

◆ Classify() [4/4]

void Classify	(	const MatType &	data,
		arma::Row< size_t > &	predictions,
		arma::rowvec &	probabilities
	)		const

Classify the given points, using this node and the entire (sub)tree beneath it.

The predicted labels for each point are returned, as well as an estimate of the probability that the prediction is correct for each point. This estimate is simply the MajorityProbability() for the leaf that each point bins to.

Parameters

data	Points to classify.
predictions	Predicted labels for each point.
probabilities	Probability estimates for each predicted label.

◆ CreateChildren()

void CreateChildren ( )

Given that this node should split, create the children.

Referenced by HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CheckInterval().

◆ MajorityClass() [1/2]

size_t MajorityClass ( ) const

inline

Get the majority class.

Definition at line 251 of file hoeffding_tree.hpp.

◆ MajorityClass() [2/2]

size_t& MajorityClass ( )

inline

Modify the majority class.

Definition at line 253 of file hoeffding_tree.hpp.

◆ MajorityProbability() [1/2]

double MajorityProbability ( ) const

inline

Get the probability of the majority class (based on training samples).

Definition at line 256 of file hoeffding_tree.hpp.

◆ MajorityProbability() [2/2]

double& MajorityProbability ( )

inline

Modify the probability of the majority class.

Definition at line 258 of file hoeffding_tree.hpp.

◆ MaxSamples() [1/2]

size_t MaxSamples ( ) const

inline

Get the maximum number of samples before a split is forced.

Definition at line 279 of file hoeffding_tree.hpp.

◆ MaxSamples() [2/2]

void MaxSamples ( const size_t maxSamples )

Modify the maximum number of samples before a split is forced.

◆ MinSamples() [1/2]

size_t MinSamples ( ) const

inline

Get the minimum number of samples for a split.

Definition at line 274 of file hoeffding_tree.hpp.

◆ MinSamples() [2/2]

void MinSamples ( const size_t minSamples )

Modify the minimum number of samples for a split.

◆ NumChildren()

size_t NumChildren ( ) const

inline

Get the number of children.

Definition at line 261 of file hoeffding_tree.hpp.

◆ NumDescendants()

size_t NumDescendants ( ) const

Get the size of the Hoeffding Tree.

Referenced by HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CheckInterval().

◆ operator=() [1/2]

HoeffdingTree& operator= ( const HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType > & other )

Copy assignment operator.

Parameters

other Tree to copy.

◆ operator=() [2/2]

HoeffdingTree& operator= ( HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType > && other )

Move assignment operator.

Parameters

other Tree to move.

◆ serialize()

void serialize	(	Archive &	ar,
		const uint32_t
	)

Serialize the split.

Referenced by HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >::CheckInterval().

◆ SplitCheck()

size_t SplitCheck ( )

Check if a split would satisfy the conditions of the Hoeffding bound with the node's specified success probability.

If so, the number of children that would be created is returned. If not, 0 is returned.

◆ SplitDimension()

size_t SplitDimension ( ) const

inline

Get the splitting dimension (size_t(-1) if no split).

Definition at line 248 of file hoeffding_tree.hpp.

◆ SuccessProbability() [1/2]

double SuccessProbability ( ) const

inline

Get the confidence required for a split.

Definition at line 269 of file hoeffding_tree.hpp.

◆ SuccessProbability() [2/2]

void SuccessProbability ( const double successProbability )

Modify the confidence required for a split.

◆ Train() [1/3]

void Train	(	const MatType &	data,
		const arma::Row< size_t > &	labels,
		const bool	batchTraining = `true`,
		const bool	resetTree = `false`,
		const size_t	numClasses = `0`
	)

Train on a set of points, either in streaming mode or in batch mode, with the given labels.

If resetTree is set to true, then reset the state of the tree to an empty tree before training.

Note that the tree will be automatically reset if the dimensionality of data does not match the dimensionality that the tree was currently trained with. The tree will also be reset if numClasses is passed.

Parameters

data	Data points to train on.
labels	Labels of data points.
batchTraining	If true, perform training in batch.
resetTree	If true, reset the tree to an empty tree before training.
numClasses	The number of classes in `labels`. Passing this will reset the tree. If not given and `resetTree` is `true`, then the number of classes will be computed from `labels`.

◆ Train() [2/3]

void Train	(	const MatType &	data,
		const data::DatasetInfo &	info,
		const arma::Row< size_t > &	labels,
		const bool	batchTraining = `true`,
		const size_t	numClasses = `0`
	)

Train on a set of points, either in streaming mode or in batch mode, with the given labels and the given DatasetInfo.

This will reset the tree. This only needs to be called when the DatasetInfo has changed—if you are training incrementally but have already passed the DatasetInfo once, use the overload of Train() that does not take a DatasetInfo and make sure resetTree is set to false.

Parameters

data	Data points to train on.
info	DatasetInfo object with information about each dimension.
labels	Labels of data points.
batchTraining	If true, perform training in batch.
numClasses	Number of classes in `labels`. If not specified, it is computed from `labels`.

◆ Train() [3/3]

void Train	(	const VecType &	point,
		const size_t	label
	)

Train on a single point in streaming mode, with the given label.

The tree will not be reset before training.

Parameters

point	Point to train on.
label	Label of point to train on.

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

/home/ryan/src/mlpack.org/_src/mlpack-git/src/mlpack/methods/hoeffding_trees/hoeffding_tree.hpp

Public Types

Public Member Functions

Detailed Description

template<typenameFitnessFunction=GiniImpurity,template<typename>classNumericSplitType=HoeffdingDoubleNumericSplit,template<typename>classCategoricalSplitType=HoeffdingCategoricalSplit> class mlpack::tree::HoeffdingTree< FitnessFunction, NumericSplitType, CategoricalSplitType >

Member Typedef Documentation

◆ CategoricalSplit

◆ NumericSplit

Constructor & Destructor Documentation

◆ HoeffdingTree() [1/5]

◆ HoeffdingTree() [2/5]

◆ HoeffdingTree() [3/5]

◆ HoeffdingTree() [4/5]

◆ HoeffdingTree() [5/5]

◆ ~HoeffdingTree()

Member Function Documentation

◆ CalculateDirection()

◆ CheckInterval() [1/2]

◆ CheckInterval() [2/2]

◆ Child() [1/2]

◆ Child() [2/2]

◆ Classify() [1/4]

◆ Classify() [2/4]

◆ Classify() [3/4]

◆ Classify() [4/4]

◆ CreateChildren()

◆ MajorityClass() [1/2]

◆ MajorityClass() [2/2]

◆ MajorityProbability() [1/2]

◆ MajorityProbability() [2/2]

◆ MaxSamples() [1/2]

◆ MaxSamples() [2/2]

◆ MinSamples() [1/2]

◆ MinSamples() [2/2]

◆ NumChildren()

◆ NumDescendants()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ serialize()

◆ SplitCheck()

◆ SplitDimension()

◆ SuccessProbability() [1/2]

◆ SuccessProbability() [2/2]

◆ Train() [1/3]

◆ Train() [2/3]

◆ Train() [3/3]