Package weka.classifiers

Interface IterativeClassifier

All Superinterfaces:

Classifier

All Known Implementing Classes:

AdaBoostM1, AdditiveRegression, FilteredClassifier, LogitBoost, MultilayerPerceptron, RandomizableFilteredClassifier

public interface IterativeClassifier
extends Classifier

Interface for classifiers that can induce models of growing complexity one step at a time.

Version:

$Revision: 15022 $

Author:

Gabi Schmidberger (gabi@cs.waikato.ac.nz), Bernhard Pfahringer (bernhard@cs.waikato.ac.nz)

Method Summary

Modifier and Type	Method	Description
void	done()	Signal end of iterating, for either the time being or permanently if setFinalized(true) has been called.
boolean	getResume()	Returns true if the classifier will be able to be trained further (with more iterations) at a later date.
void	initializeClassifier(Instances instances)	Initializes an iterative classifier.
boolean	next()	Performs one iteration.
void	setResume(boolean resume)	If called with argument true then the classifier will be able to be trained further (with more iterations) at a later date.

Methods inherited from interface weka.classifiers.Classifier

buildClassifier, classifyInstance, distributionForInstance, getCapabilities

Method Detail

initializeClassifier

void initializeClassifier(Instances instances)
                   throws java.lang.Exception

Initializes an iterative classifier.

Parameters:

instances - the instances to be used in induction

Throws:

java.lang.Exception - if the model cannot be initialized

setResume

void setResume(boolean resume)
        throws java.lang.Exception

If called with argument true then the classifier will be able to be trained further (with more iterations) at a later date. Note that this can potentially make serialized model structures larger, because certain data and data structures might need to be retained to facilitate continued training. When set to false, and done() is called, then the classifier will be cleaned up, resulting in (potentially) a more compact serialized model.

Parameters:

resume - true to enable training to continue at a later point after the initial model is built.

Throws:

java.lang.Exception - if finalization cannot be set

getResume

boolean getResume()

Returns true if the classifier will be able to be trained further (with more iterations) at a later date. Note that this can potentially make serialized model structures larger, because certain data and data structures might need to be retained to facilitate continued training. When set to false, and done() is called, then the classifier will be cleaned up, resulting in (potentially) a more compact serialized model.

Returns:

true to enable training to continue at a later point after the initial model is built.

Throws:

java.lang.Exception - if finalization cannot be set

next

boolean next()
      throws java.lang.Exception

Performs one iteration.

Returns:

false if no further iterations could be performed, true otherwise

Throws:

java.lang.Exception - if this iteration fails for unexpected reasons

done

void done()
   throws java.lang.Exception

Signal end of iterating, for either the time being or permanently if setFinalized(true) has been called. Useful for any house-keeping/cleanup.

Throws:

java.lang.Exception - if cleanup fails