edu.cmu.minorthird.classify.multi
Class MultiClassifierTeacher

java.lang.Object
  edu.cmu.minorthird.classify.multi.MultiClassifierTeacher

Direct Known Subclasses:

MultiDatasetClassifierTeacher

public abstract class MultiClassifierTeacher
extends java.lang.Object

Interface for something that trains multi label classifiers.

Author:

Cameron Williams

Constructor Summary

MultiClassifierTeacher()

Method Summary

protected abstract java.util.Iterator<MultiExample>	examplePool() Labeled instances that will be sent to the multiLearner via a call to addExample().
protected abstract boolean	hasAnswers() Return true if this teacher can answer more queries.
protected abstract java.util.Iterator<Instance>	instancePool() Unlabeled instances, which will be provided to the multiLearner via setInstancePool().
protected abstract MultiExample	labelInstance(Instance query) Label an Instance chosen by the multiLearner.
protected abstract MultiExampleSchema	schema() The set of classes that will be used.
MultiClassifier	train(ClassifierLearner learner) Train a ClassifierLearner and return the learned Classifier, using some unspecified source of information to get labels.

Methods inherited from class java.lang.Object

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

Constructor Detail

MultiClassifierTeacher

public MultiClassifierTeacher()

Method Detail

train

public final MultiClassifier train(ClassifierLearner learner)

Train a ClassifierLearner and return the learned Classifier, using some unspecified source of information to get labels.

schema

protected abstract MultiExampleSchema schema()

The set of classes that will be used.

examplePool

protected abstract java.util.Iterator<MultiExample> examplePool()

Labeled instances that will be sent to the multiLearner via a call to addExample().

instancePool

protected abstract java.util.Iterator<Instance> instancePool()

Unlabeled instances, which will be provided to the multiLearner via setInstancePool(). These can be used for semi-supervised multiLearner, or to form queries for active learning. .

labelInstance

protected abstract MultiExample labelInstance(Instance query)

Label an Instance chosen by the multiLearner. Return null if the query can't be answered, otherwise return a labeled version of the instance (an Example).

hasAnswers

protected abstract boolean hasAnswers()

Return true if this teacher can answer more queries.