Skip to content

[source]

Classification Tree#

A binary tree-based learner that greedily constructs a decision map for classification that minimizes the Gini impurity among the training labels within the leaf nodes. The height and bushiness of the tree can be determined by the user-defined max height and max leaf size hyper-parameters. Classification Trees also serve as the base learner of ensemble methods such as Random Forest and AdaBoost.

Interfaces: Estimator, Learner, Probabilistic, Ranks Features, Persistable

Data Type Compatibility: Categorical, Continuous

Parameters#

# Name Default Type Description
1 maxHeight PHP_INT_MAX int The maximum height of the tree.
2 maxLeafSize 3 int The max number of samples that a leaf node can contain.
3 minPurityIncrease 1e-7 float The minimum increase in purity necessary to continue splitting a subtree.
4 maxFeatures Auto int The max number of feature columns to consider when determining a best split.
5 maxBins Auto int The maximum number of bins to consider when determining a split with a continuous feature as the split point.

Example#

use Rubix\ML\Classifiers\ClassificationTree;

$estimator = new ClassificationTree(10, 5, 0.001, null, null);

Additional Methods#

Return the number of levels in the tree.

public height() : ?int

Return a factor that quantifies the skewness of the distribution of nodes in the tree.

public balance() : ?int

References:#


  1. W. Y. Loh. (2011). Classification and Regression Trees. 

  2. K. Alsabti. et al. (1998). CLOUDS: A Decision Tree Classifier for Large Datasets. 


Last update: 2021-10-31