ml::DecisionTrees::Node< Y > Struct Template Reference

Tree node. Nodes are split (non-terminal) or leaf (terminal). More...

#include <DecisionTreeNodes.hpp>

Inheritance diagram for ml::DecisionTrees::Node< Y >:

Public Types
typedef Eigen::Ref< const Eigen::VectorXd >	arg_type

Public Member Functions
	Node (double n_error, Y n_value, SplitNode< Y > *n_parent)
	Constructor. More...
virtual	~Node ()
	Virtual destructor.
virtual Y	operator() (arg_type x) const =0
	Returns a prediction given a feature vector. More...
virtual unsigned int	count_lower_nodes () const =0
	Total number of nodes reachable from this one.
virtual unsigned int	count_leaf_nodes () const =0
	Total number of leaf nodes reachable from this one, including itself.
virtual double	total_leaf_error () const =0
	Total error of the training samples seen by the leaf nodes reachable from this node (including its own if leaf). More...
virtual Node *	clone (SplitNode< Y > *cloned_parent) const =0
	Make a perfect copy of the node. Function works recursively from root to leafs. More...
virtual bool	is_leaf () const =0
	Return true if node is a leaf.
virtual void	collect_lowest_split_nodes (std::unordered_set< SplitNode< Y > * > &s)=0
	Adds all lowest split nodes. More...

Public Attributes
double	error
Y	value
SplitNode< Y > *	parent

Detailed Description

template<class Y>
struct ml::DecisionTrees::Node< Y >

Tree node. Nodes are split (non-terminal) or leaf (terminal).

Template Parameters

Y	Type of predicted value (integer for classification, real for regression).

See also

DecisionTree.

Member Typedef Documentation

arg_type

template<class Y >

typedef Eigen::Ref<const Eigen::VectorXd> ml::DecisionTrees::Node< Y >::arg_type

Type for feature vector.

Constructor & Destructor Documentation

Node()

template<class Y >

ml::DecisionTrees::Node< Y >::Node ( double  n_error, Y  n_value, SplitNode< Y > *  n_parent  )

inline

Constructor.

Parameters

[in]	n_error	Prediction error value on training data which reached this node.
[in]	n_value	Prediction value assigned to this node.
[in]	n_parent	Link to parent node.

Exceptions

std::domain_error

If n_error < 0.

Member Function Documentation

operator()()

template<class Y >

virtual Y ml::DecisionTrees::Node< Y >::operator() ( arg_type  x ) const

pure virtual

Returns a prediction given a feature vector.

Parameters

x	Feature vector.

Returns

Prediction value.

Implemented in ml::DecisionTrees::LeafNode< Y >.

total_leaf_error()

template<class Y >

virtual double ml::DecisionTrees::Node< Y >::total_leaf_error ( ) const

pure virtual

Total error of the training samples seen by the leaf nodes reachable from this node (including its own if leaf).

Has the invariant total_leaf_error() <= error.

Implemented in ml::DecisionTrees::LeafNode< Y >.

clone()

template<class Y >

virtual Node* ml::DecisionTrees::Node< Y >::clone ( SplitNode< Y > *  cloned_parent ) const

pure virtual

Make a perfect copy of the node. Function works recursively from root to leafs.

Parameters

[in]

cloned_parent

Pointer to already cloned parent.

Implemented in ml::DecisionTrees::LeafNode< Y >.

collect_lowest_split_nodes()

template<class Y >

virtual void ml::DecisionTrees::Node< Y >::collect_lowest_split_nodes ( std::unordered_set< SplitNode< Y > * > &  s )

pure virtual

Adds all lowest split nodes.

Walks over this node and all below it, adding to s every split node which has only leaf nodes as children.

Parameters

[out]

s

Set of pointers to split nodes.

Implemented in ml::DecisionTrees::LeafNode< Y >.

Member Data Documentation

error

template<class Y >

double ml::DecisionTrees::Node< Y >::error

Error of the training sample seen by this node, used for pruning.

value

template<class Y >

Y ml::DecisionTrees::Node< Y >::value

Value which should be returned if we stop splitting at this node.

parent

template<class Y >

SplitNode<Y>* ml::DecisionTrees::Node< Y >::parent

Link to parent node.

---

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

• ML/DecisionTreeNodes.hpp