moa.classifiers.trees

Class HoeffdingOptionTree

java.lang.Object

moa.AbstractMOAObject

moa.options.AbstractOptionHandler

moa.classifiers.AbstractClassifier

moa.classifiers.trees.HoeffdingOptionTree

All Implemented Interfaces:

Serializable, Classifier, AWTRenderable, MOAObject, OptionHandler

Direct Known Subclasses:

AdaHoeffdingOptionTree

public class HoeffdingOptionTree
extends AbstractClassifier

Hoeffding Option Tree.

Hoeffding Option Trees are regular Hoeffding trees containing additional option nodes that allow several tests to be applied, leading to multiple Hoeffding trees as separate paths. They consist of a single structure that efficiently represents multiple trees. A particular example can travel down multiple paths of the tree, contributing, in different ways, to different options.

See for details:

B. Pfahringer, G. Holmes, and R. Kirkby. New options for hoeffding trees. In AI, pages 90–99, 2007.

Parameters:

• -o : Maximum number of option paths per node
• -m : Maximum memory consumed by the tree
• -n : Numeric estimator to use :
• Gaussian approximation evaluating 10 splitpoints
• Gaussian approximation evaluating 100 splitpoints
• Greenwald-Khanna quantile summary with 10 tuples
• Greenwald-Khanna quantile summary with 100 tuples
• Greenwald-Khanna quantile summary with 1000 tuples
• VFML method with 10 bins
• VFML method with 100 bins
• VFML method with 1000 bins
• Exhaustive binary tree
• -e : How many instances between memory consumption checks
• -g : The number of instances a leaf should observe between split attempts
• -s : Split criterion to use. Example : InfoGainSplitCriterion
• -c : The allowable error in split decision, values closer to 0 will take longer to decide
• -w : The allowable error in secondary split decisions, values closer to 0 will take longer to decide
• -t : Threshold below which a split will be forced to break ties
• -b : Only allow binary splits
• -z : Memory strategy to use
• -r : Disable poor attributes
• -p : Disable pre-pruning
• -d : File to append option table to.
• -l : Leaf prediction to use: MajorityClass (MC), Naive Bayes (NB) or NaiveBayes adaptive (NBAdaptive).
• -q : The number of instances a leaf should observe before permitting Naive Bayes

Version:

$Revision: 7 $

Author:

Richard Kirkby (rkirkby@cs.waikato.ac.nz)

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	HoeffdingOptionTree.ActiveLearningNode
static class	HoeffdingOptionTree.FoundNode
static class	HoeffdingOptionTree.InactiveLearningNode
static class	HoeffdingOptionTree.LearningNode
static class	HoeffdingOptionTree.LearningNodeNB
static class	HoeffdingOptionTree.LearningNodeNBAdaptive
static class	HoeffdingOptionTree.Node
static class	HoeffdingOptionTree.SplitNode

Field Summary

Fields

Modifier and Type	Field and Description
protected double	activeLeafByteSizeEstimate
protected int	activeLeafNodeCount
FlagOption	binarySplitsOption
protected double	byteSizeEstimateOverheadFraction
protected int	decisionNodeCount
FileOption	dumpFileOption
IntOption	gracePeriodOption
protected double	inactiveLeafByteSizeEstimate
protected int	inactiveLeafNodeCount
MultiChoiceOption	leafpredictionOption
IntOption	maxByteSizeOption
IntOption	maxOptionPathsOption
protected int	maxPredictionPaths
IntOption	memoryEstimatePeriodOption
IntOption	memoryStrategyOption
IntOption	nbThresholdOption
ClassOption	nominalEstimatorOption
FlagOption	noPrePruneOption
ClassOption	numericEstimatorOption
FlagOption	removePoorAttsOption
FloatOption	secondarySplitConfidenceOption
FloatOption	splitConfidenceOption
ClassOption	splitCriterionOption
FloatOption	tieThresholdOption
protected HoeffdingOptionTree.Node	treeRoot

Fields inherited from class moa.classifiers.AbstractClassifier

classifierRandom, modelContext, randomSeed, randomSeedOption, trainingWeightSeenByModel

Fields inherited from class moa.options.AbstractOptionHandler

classOptionNamesToPreparedObjects, options

Constructor Summary

Constructors

Constructor and Description
HoeffdingOptionTree()

Method Summary

Methods

Modifier and Type	Method and Description
protected void	activateLearningNode(HoeffdingOptionTree.InactiveLearningNode toActivate, HoeffdingOptionTree.SplitNode parent, int parentBranch)
protected void	attemptToSplit(HoeffdingOptionTree.ActiveLearningNode node, HoeffdingOptionTree.SplitNode parent, int parentIndex)
int	calcByteSize()
static double	computeHoeffdingBound(double range, double confidence, double n)
void	deactivateAllLeaves()
protected void	deactivateLearningNode(HoeffdingOptionTree.ActiveLearningNode toDeactivate, HoeffdingOptionTree.SplitNode parent, int parentBranch)
void	enforceTrackerLimit()
void	estimateModelByteSizes()
protected HoeffdingOptionTree.FoundNode[]	findLearningNodes()
protected void	findLearningNodes(HoeffdingOptionTree.Node node, HoeffdingOptionTree.SplitNode parent, int parentBranch, List<HoeffdingOptionTree.FoundNode> found)
void	getModelDescription(StringBuilder out, int indent) Returns a string representation of the model.
protected Measurement[]	getModelMeasurementsImpl() Gets the current measurements of this classifier. The reason for ...Impl methods: ease programmer burden by not requiring them to remember calls to super in overridden methods.
String	getPurposeString() Gets the purpose of this object
double[]	getVotesForInstance(weka.core.Instance inst) Predicts the class memberships for a given instance.
boolean	isRandomizable() Gets whether this classifier needs a random seed.
int	measureByteSize() Gets the memory size of this object.
int	measureTreeDepth()
protected HoeffdingOptionTree.LearningNode	newLearningNode()
protected HoeffdingOptionTree.LearningNode	newLearningNode(double[] initialClassObservations)
protected AttributeClassObserver	newNominalClassObserver()
protected AttributeClassObserver	newNumericClassObserver()
void	resetLearningImpl() Resets this classifier.
void	trainOnInstanceImpl(weka.core.Instance inst) Trains this classifier incrementally using the given instance. The reason for ...Impl methods: ease programmer burden by not requiring them to remember calls to super in overridden methods.

Methods inherited from class moa.classifiers.AbstractClassifier

contextIsCompatible, copy, correctlyClassifies, getAttributeNameString, getAWTRenderer, getClassLabelString, getClassNameString, getDescription, getModelContext, getModelMeasurements, getNominalValueString, getSubClassifiers, modelAttIndexToInstanceAttIndex, modelAttIndexToInstanceAttIndex, prepareForUseImpl, resetLearning, setModelContext, setRandomSeed, trainingHasStarted, trainingWeightSeenByModel, trainOnInstance

Methods inherited from class moa.options.AbstractOptionHandler

discoverOptionsViaReflection, getCLICreationString, getOptions, getPreparedClassOption, getPreparedClassOption, prepareClassOptions, prepareForUse, prepareForUse

Methods inherited from class moa.AbstractMOAObject

copy, measureByteSize, toString

Methods inherited from class java.lang.Object

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

Methods inherited from interface moa.options.OptionHandler

getCLICreationString, getOptions, prepareForUse, prepareForUse

Field Detail

maxOptionPathsOption

public IntOption maxOptionPathsOption

maxByteSizeOption

public IntOption maxByteSizeOption

numericEstimatorOption

public ClassOption numericEstimatorOption

nominalEstimatorOption

public ClassOption nominalEstimatorOption

memoryEstimatePeriodOption

public IntOption memoryEstimatePeriodOption

gracePeriodOption

public IntOption gracePeriodOption

splitCriterionOption

public ClassOption splitCriterionOption

splitConfidenceOption

public FloatOption splitConfidenceOption

secondarySplitConfidenceOption

public FloatOption secondarySplitConfidenceOption

tieThresholdOption

public FloatOption tieThresholdOption

binarySplitsOption

public FlagOption binarySplitsOption

removePoorAttsOption

public FlagOption removePoorAttsOption

noPrePruneOption

public FlagOption noPrePruneOption

dumpFileOption

public FileOption dumpFileOption

memoryStrategyOption

public IntOption memoryStrategyOption

treeRoot

protected HoeffdingOptionTree.Node treeRoot

decisionNodeCount

protected int decisionNodeCount

activeLeafNodeCount

protected int activeLeafNodeCount

inactiveLeafNodeCount

protected int inactiveLeafNodeCount

inactiveLeafByteSizeEstimate

protected double inactiveLeafByteSizeEstimate

activeLeafByteSizeEstimate

protected double activeLeafByteSizeEstimate

byteSizeEstimateOverheadFraction

protected double byteSizeEstimateOverheadFraction

maxPredictionPaths

protected int maxPredictionPaths

leafpredictionOption

public MultiChoiceOption leafpredictionOption

nbThresholdOption

public IntOption nbThresholdOption

Constructor Detail

HoeffdingOptionTree

public HoeffdingOptionTree()

Method Detail

getPurposeString

public String getPurposeString()

Description copied from interface: OptionHandler

Gets the purpose of this object

Specified by:

getPurposeString in interface OptionHandler

Overrides:

getPurposeString in class AbstractClassifier

Returns:

the string with the purpose of this object

calcByteSize

public int calcByteSize()

measureByteSize

public int measureByteSize()

Description copied from interface: MOAObject

Gets the memory size of this object.

Specified by:

measureByteSize in interface MOAObject

Overrides:

measureByteSize in class AbstractMOAObject

Returns:

the memory size of this object

resetLearningImpl

public void resetLearningImpl()

Description copied from class: AbstractClassifier

Resets this classifier. It must be similar to starting a new classifier from scratch.

The reason for ...Impl methods: ease programmer burden by not requiring them to remember calls to super in overridden methods. Note that this will produce compiler errors if not overridden.

Specified by:

resetLearningImpl in class AbstractClassifier

trainOnInstanceImpl

public void trainOnInstanceImpl(weka.core.Instance inst)

Description copied from class: AbstractClassifier

Trains this classifier incrementally using the given instance.

The reason for ...Impl methods: ease programmer burden by not requiring them to remember calls to super in overridden methods. Note that this will produce compiler errors if not overridden.

Specified by:

trainOnInstanceImpl in class AbstractClassifier

Parameters:

inst - the instance to be used for training

getVotesForInstance

public double[] getVotesForInstance(weka.core.Instance inst)

Description copied from interface: Classifier

Predicts the class memberships for a given instance. If an instance is unclassified, the returned array elements must be all zero.

Parameters:

inst - the instance to be classified

Returns:

an array containing the estimated membership probabilities of the test instance in each class

getModelMeasurementsImpl

protected Measurement[] getModelMeasurementsImpl()

Description copied from class: AbstractClassifier

Gets the current measurements of this classifier.

The reason for ...Impl methods: ease programmer burden by not requiring them to remember calls to super in overridden methods. Note that this will produce compiler errors if not overridden.

Specified by:

getModelMeasurementsImpl in class AbstractClassifier

Returns:

an array of measurements to be used in evaluation tasks

measureTreeDepth

public int measureTreeDepth()

getModelDescription

public void getModelDescription(StringBuilder out,
                       int indent)

Description copied from class: AbstractClassifier

Returns a string representation of the model.

Specified by:

getModelDescription in class AbstractClassifier

Parameters:

out - the stringbuilder to add the description

indent - the number of characters to indent

isRandomizable

public boolean isRandomizable()

Description copied from interface: Classifier

Gets whether this classifier needs a random seed. Examples of methods that needs a random seed are bagging and boosting.

Returns:

true if the classifier needs a random seed.

computeHoeffdingBound

public static double computeHoeffdingBound(double range,
                           double confidence,
                           double n)

newNominalClassObserver

protected AttributeClassObserver newNominalClassObserver()

newNumericClassObserver

protected AttributeClassObserver newNumericClassObserver()

attemptToSplit

protected void attemptToSplit(HoeffdingOptionTree.ActiveLearningNode node,
                  HoeffdingOptionTree.SplitNode parent,
                  int parentIndex)

enforceTrackerLimit

public void enforceTrackerLimit()

estimateModelByteSizes

public void estimateModelByteSizes()

deactivateAllLeaves

public void deactivateAllLeaves()

deactivateLearningNode

protected void deactivateLearningNode(HoeffdingOptionTree.ActiveLearningNode toDeactivate,
                          HoeffdingOptionTree.SplitNode parent,
                          int parentBranch)

activateLearningNode

protected void activateLearningNode(HoeffdingOptionTree.InactiveLearningNode toActivate,
                        HoeffdingOptionTree.SplitNode parent,
                        int parentBranch)

findLearningNodes

protected HoeffdingOptionTree.FoundNode[] findLearningNodes()

findLearningNodes

protected void findLearningNodes(HoeffdingOptionTree.Node node,
                     HoeffdingOptionTree.SplitNode parent,
                     int parentBranch,
                     List<HoeffdingOptionTree.FoundNode> found)

newLearningNode

protected HoeffdingOptionTree.LearningNode newLearningNode()

newLearningNode

protected HoeffdingOptionTree.LearningNode newLearningNode(double[] initialClassObservations)