gov.sandia.cognition.learning.algorithm.ensemble

Class AdaBoost<InputType>

java.lang.Object

gov.sandia.cognition.util.AbstractCloneableSerializable

gov.sandia.cognition.algorithm.AbstractIterativeAlgorithm

gov.sandia.cognition.algorithm.AbstractAnytimeAlgorithm<ResultType>

gov.sandia.cognition.learning.algorithm.AbstractAnytimeBatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,OutputType>>,ResultType>

gov.sandia.cognition.learning.algorithm.AbstractAnytimeSupervisedBatchLearner<InputType,java.lang.Boolean,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>

gov.sandia.cognition.learning.algorithm.ensemble.AdaBoost<InputType>

Type Parameters:

InputType - The algorithm operates on a Collection<InputOutputPair<InputType,Boolean>>. The InputType will be something like Vector or String.

All Implemented Interfaces:

AnytimeAlgorithm<WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>, IterativeAlgorithm, StoppableAlgorithm, AnytimeBatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>, BatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>, BatchLearnerContainer<BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>>, SupervisedBatchLearner<InputType,java.lang.Boolean,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>, CloneableSerializable, java.io.Serializable, java.lang.Cloneable

@CodeReview(reviewer="Kevin R. Dixon",
            date="2008-07-23",
            changesNeeded=false,
            comments={"Added PublicationReference","Cleaned up javadoc a little bit with code annotations.","Otherwise, looks fine."})
 @PublicationReference(author={"Yoav Freund","Robert E.Schapire"},
                      title="A decision-theoretic generalization of on-line learning and an application to boosting",
                      publication="Journal of Computer and System Sciences",
                      notes="Volume 55, Number 1",
                      year=1997,
                      pages={119,139},
                      type=Journal,
                      url="http://www.cse.ucsd.edu/~yfreund/papers/adaboost.pdf")
public class AdaBoost<InputType>
extends AbstractAnytimeSupervisedBatchLearner<InputType,java.lang.Boolean,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>
implements BatchLearnerContainer<BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>>

The AdaBoost class implements the Adaptive Boosting (AdaBoost) algorithm formulated by Yoav Freund and Robert Shapire. It creates a weighted binary ensemble as output from the algorithm.

Since:

2.0

Author:

Justin Basilico

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
static int	DEFAULT_MAX_ITERATIONS The default maximum number of iterations is 100.
protected WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>	ensemble The ensemble learned by the algorithm.
protected BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>	weakLearner The "weak learner" that must learn from the weighted input-output pairs on each iteration.
protected java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>>	weightedData An array list containing the weighted version of the data.

Fields inherited from class gov.sandia.cognition.learning.algorithm.AbstractAnytimeBatchLearner

data, keepGoing

Fields inherited from class gov.sandia.cognition.algorithm.AbstractAnytimeAlgorithm

maxIterations

Fields inherited from class gov.sandia.cognition.algorithm.AbstractIterativeAlgorithm

DEFAULT_ITERATION, iteration

Constructor Summary

Constructors

Constructor and Description
AdaBoost() Creates a new instance of AdaBoost with no base learner.
AdaBoost(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner) Creates a new instance of AdaBoost.
AdaBoost(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner, int maxIterations) Creates a new instance of AdaBoost.

Method Summary

Modifier and Type	Method and Description
protected void	cleanupAlgorithm() Called to clean up the learning algorithm's state after learning has finished.
WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>	getEnsemble() Gets the ensemble created by this learner.
BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>	getLearner() Gets the learner contained in this object.
WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>	getResult() Gets the current result of the algorithm.
BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>	getWeakLearner() Gets the weak learner that is passed the weighted training data on each step of the algorithm.
java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>>	getWeightedData() Gets the weighted version of the data.
protected boolean	initializeAlgorithm() Called to initialize the learning algorithm's state based on the data that is stored in the data field.
protected void	setEnsemble(WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>> ensemble) Sets the ensemble created by this learner.
void	setWeakLearner(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner) Sets the weak learner that is passed the weighted training data on each step of the algorithm.
protected void	setWeightedData(java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>> weightedData) Sets the weighted version of the data.
protected boolean	step() Called to take a single step of the learning algorithm.

Methods inherited from class gov.sandia.cognition.learning.algorithm.AbstractAnytimeBatchLearner

clone, getData, getKeepGoing, learn, setData, setKeepGoing, stop

Methods inherited from class gov.sandia.cognition.algorithm.AbstractAnytimeAlgorithm

getMaxIterations, isResultValid, setMaxIterations

Methods inherited from class gov.sandia.cognition.algorithm.AbstractIterativeAlgorithm

addIterativeAlgorithmListener, fireAlgorithmEnded, fireAlgorithmStarted, fireStepEnded, fireStepStarted, getIteration, getListeners, removeIterativeAlgorithmListener, setIteration, setListeners

Methods inherited from class java.lang.Object

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

Methods inherited from interface gov.sandia.cognition.learning.algorithm.BatchLearner

learn

Methods inherited from interface gov.sandia.cognition.util.CloneableSerializable

clone

Methods inherited from interface gov.sandia.cognition.algorithm.AnytimeAlgorithm

getMaxIterations, setMaxIterations

Methods inherited from interface gov.sandia.cognition.algorithm.IterativeAlgorithm

addIterativeAlgorithmListener, getIteration, removeIterativeAlgorithmListener

Methods inherited from interface gov.sandia.cognition.algorithm.StoppableAlgorithm

isResultValid

Field Detail

DEFAULT_MAX_ITERATIONS

public static final int DEFAULT_MAX_ITERATIONS

The default maximum number of iterations is 100.

See Also:

Constant Field Values

weakLearner

protected BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner

The "weak learner" that must learn from the weighted input-output pairs on each iteration.

ensemble

protected transient WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>> ensemble

The ensemble learned by the algorithm.

weightedData

protected transient java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>> weightedData

An array list containing the weighted version of the data.

Constructor Detail

AdaBoost

public AdaBoost()

Creates a new instance of AdaBoost with no base learner.

AdaBoost

public AdaBoost(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner)

Creates a new instance of AdaBoost.

Parameters:

weakLearner - The weak learner to apply learning to on each iteration.

AdaBoost

public AdaBoost(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner,
                int maxIterations)

Creates a new instance of AdaBoost.

Parameters:

weakLearner - The weak learner to apply learning to on each iteration.

maxIterations - The maximum number of iterations to run for, which is also the upper bound on the number of learners to create.

Method Detail

initializeAlgorithm

protected boolean initializeAlgorithm()

Description copied from class: AbstractAnytimeBatchLearner

Called to initialize the learning algorithm's state based on the data that is stored in the data field. The return value indicates if the algorithm can be run or not based on the initialization.

Specified by:

initializeAlgorithm in class AbstractAnytimeBatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>

Returns:

True if the learning algorithm can be run and false if it cannot.

step

protected boolean step()

Description copied from class: AbstractAnytimeBatchLearner

Called to take a single step of the learning algorithm.

Specified by:

step in class AbstractAnytimeBatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>

Returns:

True if another step can be taken and false it the algorithm should halt.

cleanupAlgorithm

protected void cleanupAlgorithm()

Description copied from class: AbstractAnytimeBatchLearner

Called to clean up the learning algorithm's state after learning has finished.

Specified by:

cleanupAlgorithm in class AbstractAnytimeBatchLearner<java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>

getResult

public WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>> getResult()

Description copied from interface: AnytimeAlgorithm

Gets the current result of the algorithm.

Specified by:

getResult in interface AnytimeAlgorithm<WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>>>

Returns:

Current result of the algorithm.

getLearner

public BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> getLearner()

Description copied from interface: BatchLearnerContainer

Gets the learner contained in this object.

Specified by:

getLearner in interface BatchLearnerContainer<BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>>>

Returns:

The learner contained in this object.

getWeakLearner

public BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> getWeakLearner()

Gets the weak learner that is passed the weighted training data on each step of the algorithm.

Returns:

The weak learner for the algorithm to use.

setWeakLearner

public void setWeakLearner(BatchLearner<? super java.util.Collection<? extends InputOutputPair<? extends InputType,java.lang.Boolean>>,? extends Evaluator<? super InputType,? extends java.lang.Boolean>> weakLearner)

Sets the weak learner that is passed the weighted training data on each step of the algorithm.

Parameters:

weakLearner - The weak learner for the algorithm to use.

getEnsemble

public WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>> getEnsemble()

Gets the ensemble created by this learner.

Returns:

The ensemble created by this learner.

setEnsemble

protected void setEnsemble(WeightedBinaryEnsemble<InputType,Evaluator<? super InputType,? extends java.lang.Boolean>> ensemble)

Sets the ensemble created by this learner.

Parameters:

ensemble - The ensemble created by this learner.

getWeightedData

public java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>> getWeightedData()

Gets the weighted version of the data.

Returns:

The weighted version of the data.

setWeightedData

protected void setWeightedData(java.util.ArrayList<DefaultWeightedInputOutputPair<InputType,java.lang.Boolean>> weightedData)

Sets the weighted version of the data.

Parameters:

weightedData - The weighted version of the data.