edu.illinois.cs.cogcomp.lbjava.learn

Class WekaWrapper

java.lang.Object

edu.illinois.cs.cogcomp.lbjava.classify.Classifier

edu.illinois.cs.cogcomp.lbjava.learn.Learner

edu.illinois.cs.cogcomp.lbjava.learn.WekaWrapper

All Implemented Interfaces:

Serializable, Cloneable

public class WekaWrapper
extends Learner

Translates LBJava's internal problem representation into that which can be handled by WEKA learning algorithms. This translation involves storing all examples in memory so they can be passed to WEKA at one time.

WEKA must be available on your CLASSPATH in order to use this class. WEKA source code and pre-compiled jar distributions are available at: http://www.cs.waikato.ac.nz/ml/weka/

To use this class in the with clause of a learning classifier expression, the following restrictions must be recognized:

• Feature pre-extraction must be enabled.
• No hard-coded feature generators may be referenced in the using clause.
• No array producing classifiers may be referenced in the using clause.
• The names of classifiers referenced in the using clause may not contain the underscore character ('_').
• The values produced by discrete classifiers referenced in the using clause may not contain the underscore, colon, or comma characters ('_', ':', or ' ,').

To use this class in a Java application, the following restrictions must be recognized:

• doneLearning() must be called before calls to Learner.classify(Object) can be made.
• After doneLearning() is called, Learner.learn(Object) may not be called without first calling forget().

Author:

Dan Muriello

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	WekaWrapper.Parameters Simply a container for all of WekaWrapper's configurable parameters.

Field Summary

Fields

Modifier and Type	Field and Description
protected String[]	allowableValues The label producing classifier's allowable values.
protected weka.core.FastVector	attributeInfo Information about the features this learner takes as input is parsed from an attribute string and stored here.
protected String	attributeString A string encoding of the attributes used by this learner.
protected weka.classifiers.Classifier	baseClassifier Stores the instance of the WEKA classifier which we are training; default is weka.classifiers.bayes.NaiveBayes.
static String	defaultAttributeString Default for the attributeString field.
static weka.classifiers.Classifier	defaultBaseClassifier Default for the baseClassifier field.
protected weka.classifiers.Classifier	freshClassifier Stores a fresh instance of the WEKA classifier for the purposes of forgetting.
protected weka.core.Instances	instances The main collection of Instance objects.
protected boolean	trained Indicates whether the doneLearning() method has been called and the forget() method has not yet been called.

Fields inherited from class edu.illinois.cs.cogcomp.lbjava.learn.Learner

candidates, encoding, extractor, labeler, labelLexicon, lcFilePath, lexFilePath, lexicon, lossFlag, predictions, readLexiconOnDemand

Fields inherited from class edu.illinois.cs.cogcomp.lbjava.classify.Classifier

containingPackage, name

Constructor Summary

Constructors

Constructor and Description
WekaWrapper() Empty constructor.
WekaWrapper(weka.classifiers.Classifier base) Partial constructor; attribute information must be provided before any learning can occur.
WekaWrapper(weka.classifiers.Classifier base, String attributeString) Redirecting constructor.
WekaWrapper(String n) Empty constructor.
WekaWrapper(String n, weka.classifiers.Classifier base) Partial constructor; attribute information must be provided before any learning can occur.
WekaWrapper(String n, weka.classifiers.Classifier base, String attributeString) Full Constructor.
WekaWrapper(String n, String attributeString) Default Constructor.
WekaWrapper(String n, WekaWrapper.Parameters p) Initializing constructor.
WekaWrapper(WekaWrapper.Parameters p) Initializing constructor.

Method Summary

Modifier and Type	Method and Description
String[]	allowableValues() Returns the array of allowable values that a feature returned by this classifier may take.
FeatureVector	classify(int[] exampleFeatures, double[] exampleValues) This method makes one or more decisions about a single object, returning those decisions as Features in a vector.
void	doneLearning() Indicates that the classifier is finished learning.
void	forget() Destroys the learned version of the WEKA classifier and empties the instances collection of examples.
protected double[]	getDistribution(int[] exampleFeatures, double[] exampleValues) Returns a discrete distribution of the classifier's prediction values.
String	getOutputType() This learner's output type is "mixed%".
Learner.Parameters	getParameters() Retrieves the parameters that are set in this learner.
void	initializeAttributes() Takes attributeString and initializes this wrapper's instances collection to take those attributes.
void	learn(int[] exampleFeatures, double[] exampleValues, int[] exampleLabels, double[] labelValues) Since WEKA classifiers cannot learn online, this method causes no actual learning to occur, it simply creates an Instance object from this example and adds it to a set of examples from which the classifier will be built once doneLearning() is called.
void	read(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessInputStream in) Reads the binary representation of a learner with this object's run-time type, overwriting any and all learned or manually specified parameters as well as the label lexicon but without modifying the feature lexicon.
ScoreSet	scores(int[] exampleFeatures, double[] exampleValues) Produces a set of scores indicating the degree to which each possible discrete classification value is associated with the given example object.
void	setLabeler(Classifier l) Sets the labeler.
void	setParameters(WekaWrapper.Parameters p) Sets the values of parameters that control the behavior of this learning algorithm.
void	write(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessOutputStream out) Writes the learned function's internal representation in binary form.
void	write(PrintStream out) Writes the settings of the classifier in use, and a string describing the classifier, if available.

Methods inherited from class edu.illinois.cs.cogcomp.lbjava.learn.Learner

classify, classify, classify, classify, clone, countFeatures, createPrediction, createPrediction, demandLexicon, discreteValue, discreteValue, discreteValue, doneWithRound, emptyClone, featureValue, featureValue, featureValue, getCurrentLexicon, getExampleArray, getExampleArray, getExtractor, getLabeler, getLabelLexicon, getLexicon, getLexiconDiscardCounts, getLexiconLocation, getModelLocation, getPrunedLexiconSize, initialize, learn, learn, learn, learn, read, readLabelLexicon, readLearner, readLearner, readLearner, readLearner, readLearner, readLearner, readLexicon, readLexicon, readLexiconOnDemand, readLexiconOnDemand, readModel, readModel, readParameters, realValue, realValue, realValue, save, saveLexicon, saveModel, scores, scores, scoresAugmented, setCandidates, setEncoding, setExtractor, setLabelLexicon, setLexicon, setLexiconLocation, setLexiconLocation, setLossFlag, setModelLocation, setModelLocation, setParameters, setReadLexiconOnDemand, unclone, unsetLossFlag, write, writeLexicon, writeModel, writeParameters

Methods inherited from class edu.illinois.cs.cogcomp.lbjava.classify.Classifier

classify, discreteValueArray, getCompositeChildren, getInputType, realValueArray, test, toString, valueIndexOf

Methods inherited from class java.lang.Object

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

Field Detail

defaultAttributeString

public static final String defaultAttributeString

Default for the attributeString field.

See Also:

Constant Field Values

defaultBaseClassifier

public static final weka.classifiers.Classifier defaultBaseClassifier

Default for the baseClassifier field.

attributeString

protected String attributeString

A string encoding of the attributes used by this learner.

baseClassifier

protected weka.classifiers.Classifier baseClassifier

Stores the instance of the WEKA classifier which we are training; default is weka.classifiers.bayes.NaiveBayes.

freshClassifier

protected weka.classifiers.Classifier freshClassifier

Stores a fresh instance of the WEKA classifier for the purposes of forgetting.

attributeInfo

protected weka.core.FastVector attributeInfo

Information about the features this learner takes as input is parsed from an attribute string and stored here. This information is crucial in the task of interfacing with the WEKA algorithms, and must be present before the Learner.learn(Object) method can be called.

Here is an example of a valid attribute string: nom_SimpleLabel_1,2,3,:str_First:nom_Second_a,b,c,d,r,t,:num_Third:

nom stands for "Nominal", i.e. the feature SimpleLabel was declared as discrete, and had the value list {"1","2","3"}.

str stands for "Stirng", i.e. the feature First was declared to be discrete, but was not provided with a value list. When using the WekaWrapper, it is best to provide value lists whenever possible, because very few WEKA classifiers can handle string attributes.

num stands for "Numerical", i.e. the feature Third was declared to be real.

instances

protected weka.core.Instances instances

The main collection of Instance objects.

trained

protected boolean trained

Indicates whether the doneLearning() method has been called and the forget() method has not yet been called.

allowableValues

protected String[] allowableValues

The label producing classifier's allowable values.

Constructor Detail

WekaWrapper

public WekaWrapper()

Empty constructor. Instantiates this wrapper with the default learning algorithm: weka.classifiers.bayes.NaiveBayes. Attribute information must be provided before any learning can occur.

WekaWrapper

public WekaWrapper(weka.classifiers.Classifier base)

Partial constructor; attribute information must be provided before any learning can occur.

Parameters:

base - The classifier to be used in this system.

WekaWrapper

public WekaWrapper(weka.classifiers.Classifier base,
                   String attributeString)

Redirecting constructor.

Parameters:

base - The classifier to be used in this system.

attributeString - The string describing the types of attributes example objects will have.

WekaWrapper

public WekaWrapper(WekaWrapper.Parameters p)

Initializing constructor. Sets all member variables to their associated settings in the WekaWrapper.Parameters object.

Parameters:

p - The settings of all parameters.

WekaWrapper

public WekaWrapper(String n)

Empty constructor. Instantiates this wrapper with the default learning algorithm: weka.classifiers.bayes.NaiveBayes. Attribute information must be provided before any learning can occur.

Parameters:

n - The name of the classifier.

WekaWrapper

public WekaWrapper(String n,
                   weka.classifiers.Classifier base)

Partial constructor; attribute information must be provided before any learning can occur.

Parameters:

base - The classifier to be used in this system.

WekaWrapper

public WekaWrapper(String n,
                   String attributeString)

Default Constructor. Instantiates this wrapper with the default learning algorithm: weka.classifiers.bayes.NaiveBayes.

Parameters:

n - The name of the classifier.

attributeString - The string describing the types of attributes example objects will have.

WekaWrapper

public WekaWrapper(String n,
                   WekaWrapper.Parameters p)

Initializing constructor. Sets all member variables to their associated settings in the WekaWrapper.Parameters object.

Parameters:

n - The name of the classifier.

p - The settings of all parameters.

WekaWrapper

public WekaWrapper(String n,
                   weka.classifiers.Classifier base,
                   String attributeString)

Full Constructor.

Parameters:

n - The name of the classifier

base - The classifier to be used in this system.

attributeString - The string describing the types of attributes example objects will have.

Method Detail

setParameters

public void setParameters(WekaWrapper.Parameters p)

Sets the values of parameters that control the behavior of this learning algorithm.

Parameters:

p - The parameters.

getParameters

public Learner.Parameters getParameters()

Retrieves the parameters that are set in this learner.

Overrides:

getParameters in class Learner

Returns:

An object containing all the values of the parameters that control the behavior of this learning algorithm.

getOutputType

public String getOutputType()

This learner's output type is "mixed%".

Overrides:

getOutputType in class Classifier

Returns:

A string representation of the output feature type of this classifier.

initializeAttributes

public void initializeAttributes()

Takes attributeString and initializes this wrapper's instances collection to take those attributes.

setLabeler

public void setLabeler(Classifier l)

Sets the labeler.

Overrides:

setLabeler in class Learner

Parameters:

l - A labeling classifier.

allowableValues

public String[] allowableValues()

Returns the array of allowable values that a feature returned by this classifier may take.

Overrides:

allowableValues in class Classifier

Returns:

The allowable values of this learner's labeler, or an array of length zero if the labeler has not yet been established or does not specify allowable values.

learn

public void learn(int[] exampleFeatures,
                  double[] exampleValues,
                  int[] exampleLabels,
                  double[] labelValues)

Since WEKA classifiers cannot learn online, this method causes no actual learning to occur, it simply creates an Instance object from this example and adds it to a set of examples from which the classifier will be built once doneLearning() is called.

Specified by:

learn in class Learner

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

exampleLabels - The example's label(s).

labelValues - The values of the labels.

classify

public FeatureVector classify(int[] exampleFeatures,
                              double[] exampleValues)

This method makes one or more decisions about a single object, returning those decisions as Features in a vector.

Specified by:

classify in class Learner

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

Returns:

A feature vector with a single feature containing the prediction for this example.

getDistribution

protected double[] getDistribution(int[] exampleFeatures,
                                   double[] exampleValues)

Returns a discrete distribution of the classifier's prediction values.

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

forget

public void forget()

Destroys the learned version of the WEKA classifier and empties the instances collection of examples.

Overrides:

forget in class Learner

scores

public ScoreSet scores(int[] exampleFeatures,
                       double[] exampleValues)

Produces a set of scores indicating the degree to which each possible discrete classification value is associated with the given example object.

Specified by:

scores in class Learner

Parameters:

exampleFeatures - The example's array of feature indices

exampleValues - The example's array of values

Returns:

A set of scores indicating the degree to which each possible discrete classification value is associated with the given example object.

doneLearning

public void doneLearning()

Indicates that the classifier is finished learning. This method must be called if the WEKA classifier is to learn anything. Since WEKA classifiers cannot learn online, all of the training examples must be gathered and committed to first. This method invokes the WEKA classifier's buildClassifier(Instances) method.

Overrides:

doneLearning in class Learner

write

public void write(PrintStream out)

Writes the settings of the classifier in use, and a string describing the classifier, if available.

Specified by:

write in class Learner

Parameters:

out - The output stream.

write

public void write(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessOutputStream out)

Writes the learned function's internal representation in binary form.

Overrides:

write in class Learner

Parameters:

out - The output stream.

read

public void read(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessInputStream in)

Reads the binary representation of a learner with this object's run-time type, overwriting any and all learned or manually specified parameters as well as the label lexicon but without modifying the feature lexicon.

Overrides:

read in class Learner

Parameters:

in - The input stream.