org.apache.spark.ml.tree

Interface DecisionTreeModel

All Known Implementing Classes:

DecisionTreeClassificationModel, DecisionTreeRegressionModel

public interface DecisionTreeModel

Abstraction for Decision Tree models.

TODO: Add support for predicting probabilities and raw predictions SPARK-3727

Method Summary

Modifier and Type	Method and Description
int	depth() Depth of the tree.
int	maxSplitFeatureIndex() Trace down the tree, and return the largest feature index used in any split.
int	numNodes() Number of nodes in tree, including leaf nodes.
Node	rootNode() Root of the decision tree
String	toDebugString() Full description of model
DecisionTreeModel	toOld() Convert to spark.mllib DecisionTreeModel (losing some information)
String	toString() Summary of the model

Method Detail

depth

int depth()

Depth of the tree. E.g.: Depth 0 means 1 leaf node. Depth 1 means 1 internal node and 2 leaf nodes.

Returns:

(undocumented)

maxSplitFeatureIndex

int maxSplitFeatureIndex()

Trace down the tree, and return the largest feature index used in any split.

Returns:

Max feature index used in a split, or -1 if there are no splits (single leaf node).

numNodes

int numNodes()

Number of nodes in tree, including leaf nodes.

rootNode

Node rootNode()

Root of the decision tree

toDebugString

String toDebugString()

Full description of model

toOld

DecisionTreeModel toOld()

Convert to spark.mllib DecisionTreeModel (losing some information)

toString

String toString()

Summary of the model

Overrides:

toString in class Object