Class CARTRegressionTrainer

java.lang.Object

org.tribuo.common.tree.AbstractCARTTrainer<Regressor>

org.tribuo.regression.rtree.CARTRegressionTrainer

All Implemented Interfaces:

com.oracle.labs.mlrg.olcut.config.Configurable, com.oracle.labs.mlrg.olcut.provenance.Provenancable<TrainerProvenance>, DecisionTreeTrainer<Regressor>, SparseTrainer<Regressor>, Trainer<Regressor>, WeightedExamples

public final class CARTRegressionTrainer
extends AbstractCARTTrainer<Regressor>

A Trainer that uses an approximation of the CART algorithm to build a decision tree. Trains an independent tree for each output dimension.

See:

 J. Friedman, T. Hastie, & R. Tibshirani.
 "The Elements of Statistical Learning"
 Springer 2001. PDF
 

Nested Class Summary

Nested classes/interfaces inherited from class org.tribuo.common.tree.AbstractCARTTrainer

AbstractCARTTrainer.AbstractCARTTrainerProvenance

Field Summary

Fields inherited from class org.tribuo.common.tree.AbstractCARTTrainer

fractionFeaturesInSplit, maxDepth, MIN_EXAMPLES, minChildWeight, minImpurityDecrease, rng, seed, trainInvocationCounter, useRandomSplitPoints

Fields inherited from interface org.tribuo.Trainer

DEFAULT_SEED, INCREMENT_INVOCATION_COUNT

Constructor Summary

Constructors

Constructor	Description
CARTRegressionTrainer()	Creates a CART trainer.
CARTRegressionTrainer(int maxDepth)	Creates a CART trainer.
CARTRegressionTrainer(int maxDepth, float minChildWeight, float minImpurityDecrease, float fractionFeaturesInSplit, boolean useRandomSplitPoints, RegressorImpurity impurity, long seed)	Creates a CART Trainer.
CARTRegressionTrainer(int maxDepth, float minChildWeight, float minImpurityDecrease, float fractionFeaturesInSplit, RegressorImpurity impurity, long seed)	Creates a CART Trainer.

Method Summary

Modifier and Type	Method	Description
TrainerProvenance	getProvenance()
protected AbstractTrainingNode<Regressor>	mkTrainingNode(Dataset<Regressor> examples, AbstractTrainingNode.LeafDeterminer leafDeterminer)	Makes the initial training node.
String	toString()
TreeModel<Regressor>	train(Dataset<Regressor> examples, Map<String, com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance)	Trains a sparse predictive model using the examples in the given data set.
TreeModel<Regressor>	train(Dataset<Regressor> examples, Map<String, com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance, int invocationCount)	Trains a predictive model using the examples in the given data set.

Methods inherited from class org.tribuo.common.tree.AbstractCARTTrainer

getFractionFeaturesInSplit, getInvocationCount, getMinImpurityDecrease, getUseRandomSplitPoints, postConfig, setInvocationCount, train

Methods inherited from class java.lang.Object

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

Constructor Details

CARTRegressionTrainer

public CARTRegressionTrainer(int maxDepth,
 float minChildWeight,
 float minImpurityDecrease,
 float fractionFeaturesInSplit,
 boolean useRandomSplitPoints,
 RegressorImpurity impurity,
 long seed)

Creates a CART Trainer.

Parameters:

maxDepth - maxDepth The maximum depth of the tree.

minChildWeight - minChildWeight The minimum node weight to consider it for a split.

minImpurityDecrease - The minimum decrease in impurity necessary to split a node.

fractionFeaturesInSplit - fractionFeaturesInSplit The fraction of features available in each split.

useRandomSplitPoints - Whether to choose split points for features at random.

impurity - impurity The impurity function to use to determine split quality.

seed - The RNG seed.

CARTRegressionTrainer

public CARTRegressionTrainer(int maxDepth,
 float minChildWeight,
 float minImpurityDecrease,
 float fractionFeaturesInSplit,
 RegressorImpurity impurity,
 long seed)

Creates a CART Trainer.

Computes the exact split point.

Parameters:

maxDepth - maxDepth The maximum depth of the tree.

minChildWeight - minChildWeight The minimum node weight to consider it for a split.

minImpurityDecrease - The minimum decrease in impurity necessary to split a node.

fractionFeaturesInSplit - fractionFeaturesInSplit The fraction of features available in each split.

impurity - impurity The impurity function to use to determine split quality.

seed - The RNG seed.

CARTRegressionTrainer

public CARTRegressionTrainer()

Creates a CART trainer.

Sets the impurity to the MeanSquaredError, uses all the features, computes the exact split point, and sets the minimum number of examples in a leaf to AbstractCARTTrainer.MIN_EXAMPLES.

CARTRegressionTrainer

public CARTRegressionTrainer(int maxDepth)

Creates a CART trainer.

Sets the impurity to the MeanSquaredError, uses all the features, computes the exact split point and sets the minimum number of examples in a leaf to AbstractCARTTrainer.MIN_EXAMPLES.

Parameters:

maxDepth - The maximum depth of the tree.

Method Details

mkTrainingNode

protected AbstractTrainingNode<Regressor> mkTrainingNode(Dataset<Regressor> examples,
 AbstractTrainingNode.LeafDeterminer leafDeterminer)

Description copied from class: AbstractCARTTrainer

Makes the initial training node.

Specified by:

mkTrainingNode in class AbstractCARTTrainer<Regressor>

Parameters:

examples - The dataset to use.

leafDeterminer - The leaf determination function.

Returns:

The initial training node.

train

public TreeModel<Regressor> train(Dataset<Regressor> examples,
 Map<String, com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance)

Description copied from interface: SparseTrainer

Trains a sparse predictive model using the examples in the given data set.

Specified by:

train in interface SparseTrainer<Regressor>

Specified by:

train in interface Trainer<Regressor>

Overrides:

train in class AbstractCARTTrainer<Regressor>

Parameters:

examples - the data set containing the examples.

runProvenance - Training run specific provenance (e.g., fold number).

Returns:

a predictive model that can be used to generate predictions for new examples.

train

public TreeModel<Regressor> train(Dataset<Regressor> examples,
 Map<String, com.oracle.labs.mlrg.olcut.provenance.Provenance> runProvenance,
 int invocationCount)

Description copied from interface: SparseTrainer

Trains a predictive model using the examples in the given data set.

Specified by:

train in interface SparseTrainer<Regressor>

Specified by:

train in interface Trainer<Regressor>

Overrides:

train in class AbstractCARTTrainer<Regressor>

Parameters:

examples - the data set containing the examples.

runProvenance - Training run specific provenance (e.g., fold number).

invocationCount - The state of the RNG the trainer should be set to before training

Returns:

a predictive model that can be used to generate predictions for new examples.

toString

public String toString()

Overrides:

toString in class Object

getProvenance

public TrainerProvenance getProvenance()