gov.sandia.cognition.learning.algorithm.tree

Class RandomForestFactory

java.lang.Object

gov.sandia.cognition.util.AbstractCloneableSerializable

gov.sandia.cognition.learning.algorithm.tree.RandomForestFactory

All Implemented Interfaces:

CloneableSerializable, java.io.Serializable, java.lang.Cloneable

@PublicationReference(title="Bagging Predictors",
                      author="Leo Breiman",
                      year=1996,
                      type=Journal,
                      publication="Machine Learning",
                      pages={123,140},
                      url="http://www.springerlink.com/index/L4780124W2874025.pdf")
public class RandomForestFactory
extends AbstractCloneableSerializable

A factory class for creating Random Forest learners. A random forest is a combination of using bagging to create an ensemble where the ensemble members are decision trees where each split in the tree is created by only considering a random subset of the features (random subspace method).

Since:

3.4.2

Author:

Justin Basilico

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
RandomForestFactory()

Method Summary

Modifier and Type	Method and Description
static <CategoryType> BaggingCategorizerLearner<Vector,CategoryType>	createCategorizationLearner(int ensembleSize, double baggingFraction, double dimensionsFraction, int maxTreeDepth, int minLeafSize, java.util.Random random) Creates a random forest learner for categorization outputs.
static <CategoryType> BaggingRegressionLearner<Vector>	createRegressionLearner(int ensembleSize, double baggingFraction, double dimensionsFraction, int maxTreeDepth, int minLeafSize, java.util.Random random) Creates a random forest learner for categorization outputs.

Methods inherited from class gov.sandia.cognition.util.AbstractCloneableSerializable

clone

Methods inherited from class java.lang.Object

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

Constructor Detail

RandomForestFactory

public RandomForestFactory()

Method Detail

createCategorizationLearner

public static <CategoryType> BaggingCategorizerLearner<Vector,CategoryType> createCategorizationLearner(int ensembleSize,
                                                                                                        double baggingFraction,
                                                                                                        double dimensionsFraction,
                                                                                                        int maxTreeDepth,
                                                                                                        int minLeafSize,
                                                                                                        java.util.Random random)

Creates a random forest learner for categorization outputs.

Type Parameters:

CategoryType - The type of categories.

Parameters:

ensembleSize - The size of the ensemble to learn. Must be non-negative.

baggingFraction - The percentage of the data to sample (with replacement) to train each ensemble member.

dimensionsFraction - The percentage of the dimensions to sample at each node in each tree when training in order to determine the best split point.

maxTreeDepth - The maximum allowed tree depth. Must be positive.

minLeafSize - The minimum allowed number of examples that are allowed to fall into a leaf.

random - The random number generator to use.

Returns:

A new algorithm object for learning a random forest.

createRegressionLearner

public static <CategoryType> BaggingRegressionLearner<Vector> createRegressionLearner(int ensembleSize,
                                                                                      double baggingFraction,
                                                                                      double dimensionsFraction,
                                                                                      int maxTreeDepth,
                                                                                      int minLeafSize,
                                                                                      java.util.Random random)

Creates a random forest learner for categorization outputs.

Type Parameters:

CategoryType - The type of categories.

Parameters:

ensembleSize - The size of the ensemble to learn. Must be non-negative.

baggingFraction - The percentage of the data to sample (with replacement) to train each ensemble member.

dimensionsFraction - The percentage of the dimensions to sample at each node in each tree when training in order to determine the best split point.

maxTreeDepth - The maximum allowed tree depth. Must be positive.

minLeafSize - The minimum allowed number of examples that are allowed to fall into a leaf.

random - The random number generator to use.

Returns:

A new algorithm object for learning a random forest.