org.apache.ignite.ml.tree.randomforest.data

Class NodeSplit

java.lang.Object

org.apache.ignite.ml.tree.randomforest.data.NodeSplit

All Implemented Interfaces:

Serializable

public class NodeSplit
extends Object
implements Serializable

Class represents a split point for decision tree.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
NodeSplit(int featureId, double val, double gain, double impurity) Creates an instance of NodeSplit.

Method Summary

Modifier and Type	Method and Description
void	createLeaf(TreeNode node) Convert node to leaf.
double	getGain()
double	getImpurity()
double	getVal()
List<TreeNode>	split(TreeNode node) Split node from parameter onto two children nodes.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NodeSplit

public NodeSplit(int featureId,
                 double val,
                 double gain,
                 double impurity)

Creates an instance of NodeSplit.

Parameters:

featureId - Feature id.

val - Feature split value.

gain - Gain value.

impurity - Impurity value.

Method Detail

split

public List<TreeNode> split(TreeNode node)

Split node from parameter onto two children nodes.

Parameters:

node - Node.

Returns:

List of children.

createLeaf

public void createLeaf(TreeNode node)

Convert node to leaf.

Parameters:

node - Node.

getImpurity

public double getImpurity()

getGain

public double getGain()

getVal

public double getVal()