milk.classifiers
Class MIClassifier

java.lang.Object
  |
  +--milk.classifiers.MIClassifier

All Implemented Interfaces:

java.lang.Cloneable, java.io.Serializable

Direct Known Subclasses:

DD, MDD, MIBoost, MILR, MILRARITH, MILRGEOM, MINND, MIRBFNetwork, MIWrapper, TLD, TLDSimple

public abstract class MIClassifier

extends java.lang.Object

implements java.lang.Cloneable, java.io.Serializable

Abstract classifier. All schemes for numeric or nominal prediction in Weka extend this class.

See Also:

Serialized Form

Constructor Summary

MIClassifier()

Method Summary

abstract void	buildClassifier(Exemplars data) Generates a classifier.
double	classifyExemplar(Exemplar example) Classifies the given test instance.
double[]	distributionForExemplar(Exemplar example) Predicts the class memberships for a given exemplar.
static MIClassifier	forName(java.lang.String classifierName, java.lang.String[] options) Creates a new instance of a classifier given it's class name and (optional) arguments to pass to it's setOptions method.
static MIClassifier[]	makeCopies(MIClassifier model, int num) Creates copies of the current classifier, which can then be used for boosting etc.

Methods inherited from class java.lang.Object

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

Constructor Detail

MIClassifier

public MIClassifier()

Method Detail

buildClassifier

public abstract void buildClassifier(Exemplars data)
                              throws java.lang.Exception

Generates a classifier. Must initialize all fields of the classifier that are not being set via options (ie. multiple calls of buildClassifier must always lead to the same result). Must not change the dataset in any way.

Parameters:

data - set of exemplars serving as training data

Throws:

java.lang.Exception - if the classifier has not been generated successfully

distributionForExemplar

public double[] distributionForExemplar(Exemplar example)
                                 throws java.lang.Exception

Predicts the class memberships for a given exemplar. If an instance is unclassified, the returned array elements must be all zero. If the class is numeric, the array must consist of only one element, which contains the predicted value. Classifiers have to implement either this or classifyExemplar().

Parameters:

example - the instance to be classified

Returns:

an array containing the estimated membership probabilities of the test instance in each class (this should sum to at most 1)

Throws:

java.lang.Exception - if distribution could not be computed successfully

classifyExemplar

public double classifyExemplar(Exemplar example)
                        throws java.lang.Exception

Classifies the given test instance. The instance has to belong to a dataset when it's being classified. Classifiers either have to implement this or distributionForExemplar().

Parameters:

example - the instance to be classified

Returns:

the predicted most likely class for the instance or Instance.missingValue() if no prediction is made

Throws:

java.lang.Exception - if an error occurred during the prediction

forName

public static MIClassifier forName(java.lang.String classifierName,
                                   java.lang.String[] options)
                            throws java.lang.Exception

Creates a new instance of a classifier given it's class name and (optional) arguments to pass to it's setOptions method. If the classifier implements OptionHandler and the options parameter is non-null, the classifier will have it's options set.

Parameters:

classifierName - the fully qualified class name of the classifier

options - an array of options suitable for passing to setOptions. May be null.

Returns:

the newly created MI-classifier, ready for use.

Throws:

java.lang.Exception - if the classifier name is invalid, or the options supplied are not acceptable to the classifier

makeCopies

public static MIClassifier[] makeCopies(MIClassifier model,
                                        int num)
                                 throws java.lang.Exception

Creates copies of the current classifier, which can then be used for boosting etc. Note that this method now uses Serialization to perform a deep copy, so the Classifier object must be fully Serializable. Any currently built model will now be copied as well.

Parameters:

model - an example classifier to copy

num - the number of classifiers copies to create.

Returns:

an array of MI-classifiers.

Throws:

java.lang.Exception - if an error occurs