moa.classifiers

Class AbstractClassifier

java.lang.Object

moa.AbstractMOAObject

moa.options.AbstractOptionHandler

moa.classifiers.AbstractClassifier

All Implemented Interfaces:

Serializable, Classifier, AWTRenderable, MOAObject, OptionHandler

Direct Known Subclasses:

AbstractAMRules, AccuracyUpdatedEnsemble, AccuracyWeightedEnsemble, ActiveClassifier, ChangeDetectorLearner, DACC, DecisionStump, DriftDetectionMethodClassifier, FIMTDD, HoeffdingOptionTree, HoeffdingTree, kNN, LeveragingBag, LimAttClassifier, MajorityClass, MajorityLabelset, NaiveBayes, NaiveBayesMultinomial, NoChange, OCBoost, OnlineAccuracyUpdatedEnsemble, OnlineSmoothBoost, ORTO, OzaBag, OzaBagAdwin, OzaBoost, OzaBoostAdwin, Perceptron, Perceptron, RandomRules, RuleClassifier, SGD, SGDMultiClass, SPegasos, TargetMean, TemporallyAugmentedClassifier, WeightedMajorityAlgorithm, WEKAClassifier

public abstract class AbstractClassifier
extends AbstractOptionHandler
implements Classifier

Abstract Classifier. All learners for nominal prediction in MOA extend this class.

Version:

$Revision: 7 $

Author:

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

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
Random	classifierRandom Random Generator used in randomizable learners
protected InstancesHeader	modelContext Header of the instances of the data stream
protected int	randomSeed Random seed used in randomizable learners
IntOption	randomSeedOption Option for randomizable learners to change the random seed
protected double	trainingWeightSeenByModel Sum of the weights of the instances trained by this model

Fields inherited from class moa.options.AbstractOptionHandler

classOptionNamesToPreparedObjects, options

Constructor Summary

Constructors

Constructor and Description
AbstractClassifier() Creates an classifier and setups the random seed option if the classifier is randomizable.

Method Summary

Methods

Modifier and Type	Method and Description
static boolean	contextIsCompatible(InstancesHeader originalContext, InstancesHeader newContext) Returns if two contexts or headers of instances are compatible. Two contexts are compatible if they follow the following rules: Rule 1: num classes can increase but never decrease Rule 2: num attributes can increase but never decrease Rule 3: num nominal attribute values can increase but never decrease Rule 4: attribute types must stay in the same order (although class can move; is always skipped over) Attribute names are free to change, but should always still represent the original attributes.
Classifier	copy() This method produces a copy of this object.
boolean	correctlyClassifies(weka.core.Instance inst) Gets whether this classifier correctly classifies an instance.
String	getAttributeNameString(int attIndex) Gets the name of an attribute from the header.
AWTRenderer	getAWTRenderer() Returns the AWT Renderer
String	getClassLabelString(int classLabelIndex) Gets the name of a label of the class from the header.
String	getClassNameString() Gets the name of the attribute of the class from the header.
void	getDescription(StringBuilder out, int indent) Returns a string representation of this object.
InstancesHeader	getModelContext() Gets the reference to the header of the data stream.
abstract void	getModelDescription(StringBuilder out, int indent) Returns a string representation of the model.
Measurement[]	getModelMeasurements() Gets the current measurements of this classifier.
protected abstract 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	getNominalValueString(int attIndex, int valIndex) Gets the name of a value of an attribute from the header.
String	getPurposeString() Gets the purpose of this object
Classifier[]	getSubClassifiers() Gets the classifiers of this ensemble.
protected static int	modelAttIndexToInstanceAttIndex(int index, weka.core.Instance inst) Gets the index of the attribute in the instance, given the index of the attribute in the learner.
protected static int	modelAttIndexToInstanceAttIndex(int index, weka.core.Instances insts) Gets the index of the attribute in a set of instances, given the index of the attribute in the learner.
void	prepareForUseImpl(TaskMonitor monitor, ObjectRepository repository) This method describes the implementation of how to prepare this object for use.
void	resetLearning() Resets this classifier.
abstract void	resetLearningImpl() Resets this classifier.
void	setModelContext(InstancesHeader ih) Sets the reference to the header of the data stream.
void	setRandomSeed(int s) Sets the seed for random number generation.
boolean	trainingHasStarted() Gets whether training has started.
double	trainingWeightSeenByModel() Gets the sum of the weights of the instances that have been used by this classifier during the training in trainOnInstance
void	trainOnInstance(weka.core.Instance inst) Trains this classifier incrementally using the given instance.
abstract 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.options.AbstractOptionHandler

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

Methods inherited from class moa.AbstractMOAObject

copy, measureByteSize, measureByteSize, toString

Methods inherited from class java.lang.Object

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

Methods inherited from interface moa.classifiers.Classifier

getVotesForInstance, isRandomizable

Methods inherited from interface moa.options.OptionHandler

getCLICreationString, getOptions, prepareForUse, prepareForUse

Methods inherited from interface moa.MOAObject

measureByteSize

Field Detail

modelContext

protected InstancesHeader modelContext

Header of the instances of the data stream

trainingWeightSeenByModel

protected double trainingWeightSeenByModel

Sum of the weights of the instances trained by this model

randomSeed

protected int randomSeed

Random seed used in randomizable learners

randomSeedOption

public IntOption randomSeedOption

Option for randomizable learners to change the random seed

classifierRandom

public Random classifierRandom

Random Generator used in randomizable learners

Constructor Detail

AbstractClassifier

public AbstractClassifier()

Creates an classifier and setups the random seed option if the classifier is randomizable.

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 AbstractOptionHandler

Returns:

the string with the purpose of this object

prepareForUseImpl

public void prepareForUseImpl(TaskMonitor monitor,
                     ObjectRepository repository)

Description copied from class: AbstractOptionHandler

This method describes the implementation of how to prepare this object for use. All classes that extends this class have to implement prepareForUseImpl and not prepareForUse since prepareForUse calls prepareForUseImpl.

Specified by:

prepareForUseImpl in class AbstractOptionHandler

Parameters:

monitor - the TaskMonitor to use

repository - the ObjectRepository to use

setModelContext

public void setModelContext(InstancesHeader ih)

Description copied from interface: Classifier

Sets the reference to the header of the data stream. The header of the data stream is extended from WEKA Instances. This header is needed to know the number of classes and attributes

Specified by:

setModelContext in interface Classifier

Parameters:

ih - the reference to the data stream header

getModelContext

public InstancesHeader getModelContext()

Description copied from interface: Classifier

Gets the reference to the header of the data stream. The header of the data stream is extended from WEKA Instances. This header is needed to know the number of classes and attributes

Specified by:

getModelContext in interface Classifier

Returns:

the reference to the data stream header

setRandomSeed

public void setRandomSeed(int s)

Description copied from interface: Classifier

Sets the seed for random number generation.

Specified by:

setRandomSeed in interface Classifier

Parameters:

s - the seed

trainingHasStarted

public boolean trainingHasStarted()

Description copied from interface: Classifier

Gets whether training has started.

Specified by:

trainingHasStarted in interface Classifier

Returns:

true if training has started

trainingWeightSeenByModel

public double trainingWeightSeenByModel()

Description copied from interface: Classifier

Gets the sum of the weights of the instances that have been used by this classifier during the training in trainOnInstance

Specified by:

trainingWeightSeenByModel in interface Classifier

Returns:

the weight of the instances that have been used training

resetLearning

public void resetLearning()

Description copied from interface: Classifier

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

Specified by:

resetLearning in interface Classifier

trainOnInstance

public void trainOnInstance(weka.core.Instance inst)

Description copied from interface: Classifier

Trains this classifier incrementally using the given instance.

Specified by:

trainOnInstance in interface Classifier

Parameters:

inst - the instance to be used for training

getModelMeasurements

public Measurement[] getModelMeasurements()

Description copied from interface: Classifier

Gets the current measurements of this classifier.

Specified by:

getModelMeasurements in interface Classifier

Returns:

an array of measurements to be used in evaluation tasks

getDescription

public void getDescription(StringBuilder out,
                  int indent)

Description copied from interface: MOAObject

Returns a string representation of this object. Used in AbstractMOAObject.toString to give a string representation of the object.

Specified by:

getDescription in interface MOAObject

Parameters:

out - the stringbuilder to add the description

indent - the number of characters to indent

getSubClassifiers

public Classifier[] getSubClassifiers()

Description copied from interface: Classifier

Gets the classifiers of this ensemble. Returns null if this classifier is a single classifier.

Specified by:

getSubClassifiers in interface Classifier

Returns:

an array of the classifiers of the ensemble

copy

public Classifier copy()

Description copied from interface: MOAObject

This method produces a copy of this object.

Specified by:

copy in interface Classifier

Specified by:

copy in interface MOAObject

Specified by:

copy in interface OptionHandler

Overrides:

copy in class AbstractOptionHandler

Returns:

a copy of this object

correctlyClassifies

public boolean correctlyClassifies(weka.core.Instance inst)

Description copied from interface: Classifier

Gets whether this classifier correctly classifies an instance. Uses getVotesForInstance to obtain the prediction and the instance to obtain its true class.

Specified by:

correctlyClassifies in interface Classifier

Parameters:

inst - the instance to be classified

Returns:

true if the instance is correctly classified

getClassNameString

public String getClassNameString()

Gets the name of the attribute of the class from the header.

Returns:

the string with name of the attribute of the class

getClassLabelString

public String getClassLabelString(int classLabelIndex)

Gets the name of a label of the class from the header.

Parameters:

classLabelIndex - the label index

Returns:

the name of the label of the class

getAttributeNameString

public String getAttributeNameString(int attIndex)

Gets the name of an attribute from the header.

Parameters:

attIndex - the attribute index

Returns:

the name of the attribute

getNominalValueString

public String getNominalValueString(int attIndex,
                           int valIndex)

Gets the name of a value of an attribute from the header.

Parameters:

attIndex - the attribute index

valIndex - the value of the attribute

Returns:

the name of the value of the attribute

contextIsCompatible

public static boolean contextIsCompatible(InstancesHeader originalContext,
                          InstancesHeader newContext)

Returns if two contexts or headers of instances are compatible.

Two contexts are compatible if they follow the following rules:
Rule 1: num classes can increase but never decrease
Rule 2: num attributes can increase but never decrease
Rule 3: num nominal attribute values can increase but never decrease
Rule 4: attribute types must stay in the same order (although class can move; is always skipped over)

Attribute names are free to change, but should always still represent the original attributes.

Parameters:

originalContext - the first context to compare

newContext - the second context to compare

Returns:

true if the two contexts are compatible.

getAWTRenderer

public AWTRenderer getAWTRenderer()

Returns the AWT Renderer

Specified by:

getAWTRenderer in interface AWTRenderable

Returns:

the AWT Renderer

resetLearningImpl

public abstract void resetLearningImpl()

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.

trainOnInstanceImpl

public abstract 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. Note that this will produce compiler errors if not overridden.

Parameters:

inst - the instance to be used for training

getModelMeasurementsImpl

protected abstract 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. Note that this will produce compiler errors if not overridden.

Returns:

an array of measurements to be used in evaluation tasks

getModelDescription

public abstract void getModelDescription(StringBuilder out,
                       int indent)

Returns a string representation of the model.

Parameters:

out - the stringbuilder to add the description

indent - the number of characters to indent

modelAttIndexToInstanceAttIndex

protected static int modelAttIndexToInstanceAttIndex(int index,
                                  weka.core.Instance inst)

Gets the index of the attribute in the instance, given the index of the attribute in the learner.

Parameters:

index - the index of the attribute in the learner

inst - the instance

Returns:

the index in the instance

modelAttIndexToInstanceAttIndex

protected static int modelAttIndexToInstanceAttIndex(int index,
                                  weka.core.Instances insts)

Gets the index of the attribute in a set of instances, given the index of the attribute in the learner.

Parameters:

index - the index of the attribute in the learner

insts - the instances

Returns:

the index of the attribute in the instances