de.jstacs.classifier
Class AbstractClassifier

java.lang.Object
  de.jstacs.classifier.AbstractClassifier

All Implemented Interfaces:

Storable, Cloneable

Direct Known Subclasses:

AbstractScoreBasedClassifier

public abstract class AbstractClassifier
extends Object
implements Storable, Cloneable

The super class for any classifier.

The order of the classes is never changed inside the classifier. The samples you put in the methods like train, test and evaluate should always have the same order that you have used while instantiation of the object.

For two classes it is highly recommended to set the foreground as first class and the second class as background.

Author:

Jens Keilwagen, Jan Grau

Constructor Summary

AbstractClassifier(AlphabetContainer abc)
The constructor for a homogeneous classifier.

AbstractClassifier(AlphabetContainer abc, int length)
The constructor for an inhomogeneous classifier.

AbstractClassifier(StringBuffer xml)
The constructor for the Storable interface.

Method Summary

byte[]	classify(Sample s) This method classifies all sequences of a sample and returns the index i, with 0 < i < getNumberOfClasses(), of the class to which the sequence is assigned.
abstract byte	classify(Sequence seq) This method classifies a sequence and returns the index i, with 0 < i < getNumberOfClasses(), of the class to which the sequence is assigned.
AbstractClassifier	clone()
NumericalResultSet	evaluate(MeasureParameters params, boolean exceptionIfNotComputeable, Sample... s) This method evaluates the classifier and computes all numerical results as for instance the sensitivity for a given specificity, the area under ROC curve and so on.
ResultSet	evaluateAll(MeasureParameters params, boolean exceptionIfNotComputeable, Sample... s) This method evaluates the classifier and computes all results.
protected abstract void	extractFurtherClassifierInfosFromXML(StringBuffer xml) Extracts further information of a classifier from an xml-representation.
AlphabetContainer	getAlphabetContainer() This method return the container of alphabets that is used in the classifier.
ResultSet	getCharacteristics() Returns some information characterizing or describing the current instance of the model.
protected NumericalResult	getClassificationRate(Sample[] s) This method computes the classification rate for a given array of samples.
abstract CategoricalResult[]	getClassifierAnnotation() Returns an array of Results of dimension getNumberOfClasses that contains information the classifier and for each class.
protected abstract StringBuffer	getFurtherClassifierInfos() This method returns further information of a classifier as a StringBuffer.
abstract String	getInstanceName() Returns a short description of the classifier.
int	getLength() Returns the length of the sequences this classifier can handle or 0 for sequences of arbitrary length.
static MeasureParameters	getMeasuresForEvaluate() Returns an object of the parameters for the evaluate-method.
static MeasureParameters	getMeasuresForEvaluateAll() Returns an object of the parameters for the evaluateAll-method.
abstract int	getNumberOfClasses() Returns the number of classes that can be distinguished.
abstract NumericalResultSet	getNumericalCharacteristics() Returns the subset of numerical values that are also returned by getCharacteristsics.
protected LinkedList<? extends Result>	getResults(Sample[] s, MeasureParameters params, boolean exceptionIfNotComputeable, boolean all) This method computes the results for any evaluation of the classifier.
protected abstract String	getXMLTag() Returns the String that is used as tag for the xml-representation.
abstract boolean	isTrained() This method the state of the classifier.
boolean	setNewAlphabetContainerInstance(AlphabetContainer abc) This method tries to set a new instance of an AlphabetConatiner for the current model.
ConfusionMatrix	test(Sample... testData) This method computes the confusion matrix for a given array of test data
StringBuffer	toXML() This method returns an XML-representation of an instance of the implementing class.
void	train(Sample... s) Trains the AbstractClassifier object given the data as Samples.
abstract void	train(Sample[] s, double[][] weights) This method trains a classifier over an array of weighted Samples.

Methods inherited from class java.lang.Object

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

Constructor Detail

AbstractClassifier

public AbstractClassifier(AlphabetContainer abc)

The constructor for a homogeneous classifier. Such a classifier can handle sequences of arbitrary length.

Parameters:

abc -

AbstractClassifier

public AbstractClassifier(AlphabetContainer abc,
                          int length)
                   throws IllegalArgumentException

The constructor for an inhomogeneous classifier. Such a classifier can handle sequences of arbitrary length.

Parameters:

abc - the alphabets that are used for

length - the length of the sequences that can be classified

Throws:

IllegalArgumentException - if the length and the possible length of the AlphabetContainer does not match

AbstractClassifier

public AbstractClassifier(StringBuffer xml)
                   throws NonParsableException

The constructor for the Storable interface.

Parameters:

xml - the xml-representation

Throws:

NonParsableException - if the xml-format is not parsable

Method Detail

getMeasuresForEvaluate

public static final MeasureParameters getMeasuresForEvaluate()
                                                      throws ParameterException

Returns an object of the parameters for the evaluate-method. The parameters can be set and the measures can be switch on or off.

Returns:

an object of the parameters for the evaluate-method

Throws:

ParameterException

See Also:

evaluate(MeasureParameters, boolean, Sample...)

getMeasuresForEvaluateAll

public static final MeasureParameters getMeasuresForEvaluateAll()
                                                         throws ParameterException

Returns an object of the parameters for the evaluateAll-method. The parameters can be set and the measures can be switch on or off.

Returns:

an object of the parameters for the evaluateAll-method

Throws:

ParameterException

See Also:

evaluateAll(MeasureParameters, boolean, Sample...)

classify

public abstract byte classify(Sequence seq)
                       throws Exception

This method classifies a sequence and returns the index i, with 0 < i < getNumberOfClasses(), of the class to which the sequence is assigned.

This method should check that the sequence is defined over the underlying alphabet and length.

Parameters:

seq - the sequence to be classified

Returns:

the index of the class to which the sequence is assigned

Throws:

Exception - if the classifier is not trained or something is wrong with the sequence

classify

public byte[] classify(Sample s)
                throws Exception

This method classifies all sequences of a sample and returns the index i, with 0 < i < getNumberOfClasses(), of the class to which the sequence is assigned.

Parameters:

s - the sample to be classified

Returns:

an array of class assignments

Throws:

Exception - if something went wrong while classification

clone

public AbstractClassifier clone()
                         throws CloneNotSupportedException

Overrides:

clone in class Object

Throws:

CloneNotSupportedException

evaluate

public final NumericalResultSet evaluate(MeasureParameters params,
                                         boolean exceptionIfNotComputeable,
                                         Sample... s)
                                  throws Exception

This method evaluates the classifier and computes all numerical results as for instance the sensitivity for a given specificity, the area under ROC curve and so on. This method should be used in any kind of classifier assessment as for instance crossvalidation, hold out sampling, ... .

For two classes it is highly recommended to set the foreground as first class and the second class as background, i.e. the first sample should be the foreground sample and the second should be background sample. See also this comment.

Parameters:

params - the current parameters

exceptionIfNotComputeable - if true the method

s - the array of Samples

Returns:

a set of numerical results

Throws:

Exception - if something went wrong

See Also:

ClassifierAssessment.assess(MeasureParameters, ClassifierAssessmentAssessParameterSet, ProgressUpdater, Sample[]), ClassifierAssessment.assess(MeasureParameters, ClassifierAssessmentAssessParameterSet, Sample[]), ClassifierAssessment.assess(MeasureParameters, ClassifierAssessmentAssessParameterSet, ProgressUpdater, Sample[][][])

evaluateAll

public final ResultSet evaluateAll(MeasureParameters params,
                                   boolean exceptionIfNotComputeable,
                                   Sample... s)
                            throws Exception

This method evaluates the classifier and computes all results.

Parameters:

params - the current parameters

exceptionIfNotComputeable - if true the method throws an exception if a measure could not be computed, otherwise it is ignored

s - the array of Samples

Returns:

a set of results

Throws:

Exception - if something went wrong

getResults

protected LinkedList<? extends Result> getResults(Sample[] s,
                                                  MeasureParameters params,
                                                  boolean exceptionIfNotComputeable,
                                                  boolean all)
                                           throws Exception

This method computes the results for any evaluation of the classifier.

Parameters:

s - the array of Samples

params - the current parameters

exceptionIfNotComputeable - if true the method throws an exception if a measure could not be computed, otherwise it is ignored

all - if true the method computes all results, if false it computes only the numerical results

Returns:

a list of results

Throws:

Exception - if something went wrong

See Also:

evaluate(MeasureParameters, boolean, Sample...), evaluateAll(MeasureParameters, boolean, Sample...)

getClassificationRate

protected final NumericalResult getClassificationRate(Sample[] s)
                                               throws Exception

This method computes the classification rate for a given array of samples.

Parameters:

s - the array of samples; sample 0 contains only elements of class 0; sample 1 ...

Returns:

the classification rate

Throws:

Exception - if something went wrong while classification

getAlphabetContainer

public final AlphabetContainer getAlphabetContainer()

This method return the container of alphabets that is used in the classifier.

Returns:

the used alphabet

getCharacteristics

public ResultSet getCharacteristics()
                             throws Exception

Returns some information characterizing or describing the current instance of the model. This could be for instance the number of edges for a Bayesian network or an image showing some representation of the model of a class. The set of characteristics should always include the XML-representation of the classifier. The corresponding result type is ObjectResult

Returns:

the characteristics

Throws:

Exception - an Exception is thrown if some of the characteristics could not be defined

See Also:

StorableResult

getInstanceName

public abstract String getInstanceName()

Returns a short description of the classifier.

Returns:

a short description of the classifier

getClassifierAnnotation

public abstract CategoricalResult[] getClassifierAnnotation()

Returns an array of Results of dimension getNumberOfClasses that contains information the classifier and for each class. res[0] = new CategoricalResult( "classifier", "the kind of classifier", getInstanceName() ); res[1] = new CategoricalResult( "class info 0", "some information about the class", "info0" ); res[2] = new CategoricalResult( "class info 1", "some information about the class", "info1" ); ...

Returns:

an array of Results that contains information the classifier

getLength

public final int getLength()

Returns the length of the sequences this classifier can handle or 0 for sequences of arbitrary length.

Returns:

the length

getNumericalCharacteristics

public abstract NumericalResultSet getNumericalCharacteristics()
                                                        throws Exception

Returns the subset of numerical values that are also returned by getCharacteristsics.

Returns:

the numerical characteristics

Throws:

Exception - an Exception is thrown if some of the characteristics could not be defined

getNumberOfClasses

public abstract int getNumberOfClasses()

Returns the number of classes that can be distinguished. For example if distinguish between foreground and background this method should return 2, even if you use a mixture model for either foreground or background.

Returns:

the number of classes that can be distinguished

isTrained

public abstract boolean isTrained()

This method the state of the classifier.

Returns:

true if the classifier is able to classify sequences, otherwise false

setNewAlphabetContainerInstance

public boolean setNewAlphabetContainerInstance(AlphabetContainer abc)

This method tries to set a new instance of an AlphabetConatiner for the current model. This instance has to be consistent with the underlying instance of an AlphabetContainer.

This method can be very useful to save time.

Parameters:

abc - the alphabets

Returns:

true if the new container could be set

See Also:

getAlphabetContainer(), AlphabetContainer.checkConsistency(AlphabetContainer)

test

public ConfusionMatrix test(Sample... testData)
                     throws Exception,
                            ClassDimensionException

This method computes the confusion matrix for a given array of test data

Parameters:

testData - the data

Returns:

the confusion matrix

Throws:

ClassDimensionException - if the number of samples in incorrect

Exception - if something went wrong

train

public void train(Sample... s)
           throws Exception

Trains the AbstractClassifier object given the data as Samples.
This method should work non-incrementally. That means the result of the following series: train(data1); train(data2); should be a fully trained model over data2 and not over data1, data2.

This method should check that the Samples are defined over the underlying alphabet and length.

Parameters:

s - the data

• either an array of Samples: train( new Sample[]{s1,s2,s3}) or
• a enumeration of Samples: train(s1,s2,s3)

Throws:

Exception - an Exception should be thrown if the training did not succeed

train

public abstract void train(Sample[] s,
                           double[][] weights)
                    throws Exception

This method trains a classifier over an array of weighted Samples. That's why the following has to be fulfilled:

• s.length == weights.length
• and for all i: weights[i] == null || s[i].getNumberOfElements() == weights[i].length.

This method should work non-incrementally as the method train( Sample[] ).

This method should check that the Samples are defined over the underlying alphabet and length.

Parameters:

s - an array of Samples

weights - the weights for s

Throws:

Exception - an Exception should be thrown if the weights are incorrect or the training did not succeed

getXMLTag

protected abstract String getXMLTag()

Returns the String that is used as tag for the xml-representation. This method is used by the methods fromXML( StringBuffer ) and toXML().

Returns:

the String that is used as tag for the xml-representation

extractFurtherClassifierInfosFromXML

protected abstract void extractFurtherClassifierInfosFromXML(StringBuffer xml)
                                                      throws NonParsableException

Extracts further information of a classifier from an xml-representation. This method is used by the method fromXML( StringBuffer ) and should not be made public.

Parameters:

xml - the xml-representation

Throws:

NonParsableException

toXML

public final StringBuffer toXML()

Description copied from interface: Storable

This method returns an XML-representation of an instance of the implementing class.

Specified by:

toXML in interface Storable

Returns:

the XML-representation

getFurtherClassifierInfos

protected abstract StringBuffer getFurtherClassifierInfos()

This method returns further information of a classifier as a StringBuffer. This method is used by the method toXML() and should not be made public.

Returns:

further information of a classifier as a StringBuffer