it.uniroma2.sag.kelp.data.examplegenerator

Class SequenceExampleGeneratorLinear

java.lang.Object

it.uniroma2.sag.kelp.data.examplegenerator.SequenceExampleGeneratorLinear

All Implemented Interfaces:

SequenceExampleGenerator

public class SequenceExampleGeneratorLinear
extends Object
implements SequenceExampleGenerator

A SequenceExampleGeneratorLinearAlg allows to explicitly enrich a targeted Example (reflecting an item in a SequenceExample) with information derived from the s assigned to the previous n examples.
Given a representation used to represent an example, this class generates a copy of an input Example enriched with additional features reflecting the classes assigned to the previous examples in the sequence.
This class should be used when the learning algorithm used within the SequenceClassificationLearningAlgorithm implements the LinearMethod interface.

Author:

Danilo Croce

Constructor Summary

Constructors

Constructor and Description
SequenceExampleGeneratorLinear()
SequenceExampleGeneratorLinear(int transitionsOrder, String originalRepresentationName, float transitionWeight)

Method Summary

Methods

Modifier and Type	Method and Description
Example	generateExampleWithHistory(SequenceExample sequenceExample, SequencePath p, int elementId) At labeling time, this method allows to enrich a specific Example with the labels assigned by the classifier to the previous Examples
SequenceExample	generateSequenceExampleEnrichedWithHistory(SequenceExample sequenceExample) This method allows to enrich each Example from an input SequenceExample with the labels assigned by the classifier to the previous Examples
String	getRepresentationName()
int	getTransitionsOrder()
float	getTransitionWeight()
void	setRepresentationName(String representationName)

Methods inherited from class java.lang.Object

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

Constructor Detail

SequenceExampleGeneratorLinear

public SequenceExampleGeneratorLinear()

SequenceExampleGeneratorLinear

public SequenceExampleGeneratorLinear(int transitionsOrder,
                              String originalRepresentationName,
                              float transitionWeight)

Parameters:

transitionsOrder - The number of examples preceding a target example to be considered during the manipulation process

originalRepresentationName - The identifier of the representation used to represent an example in the sequence and which will be enriched

enrichedWithHistoryRepresentationName - The identifier of the new representation produced in the manipulation process

transitionWeight - The weight to assign to each new feature added in the manipulation process

Method Detail

generateExampleWithHistory

public Example generateExampleWithHistory(SequenceExample sequenceExample,
                                 SequencePath p,
                                 int elementId)

Description copied from interface: SequenceExampleGenerator

At labeling time, this method allows to enrich a specific Example with the labels assigned by the classifier to the previous Examples

Specified by:

generateExampleWithHistory in interface SequenceExampleGenerator

Parameters:

sequenceExample - The targeted sequence

p - the sequence of Label assigned from a classifier to the SequenceExamlpe

elementId - the offset of the targeted word in the sequence

Returns:

generateSequenceExampleEnrichedWithHistory

public SequenceExample generateSequenceExampleEnrichedWithHistory(SequenceExample sequenceExample)

Description copied from interface: SequenceExampleGenerator

This method allows to enrich each Example from an input SequenceExample with the labels assigned by the classifier to the previous Examples

Specified by:

generateSequenceExampleEnrichedWithHistory in interface SequenceExampleGenerator

Parameters:

sequenceExample - The input sequence

Returns:

getRepresentationName

public String getRepresentationName()

Returns:

The identifier of the representation used to represent an example in the sequence and which will be enriched

getTransitionsOrder

public int getTransitionsOrder()

Specified by:

getTransitionsOrder in interface SequenceExampleGenerator

Returns:

the number n of elements (in the sequence) whose labels are to be considered to enrich a targeted element

getTransitionWeight

public float getTransitionWeight()

Returns:

The weight to assign to each new feature added in the manipulation process

setRepresentationName

public void setRepresentationName(String representationName)

Parameters:

representationName - The identifier of the representation used to represent a example in the sequence and which will be enriched