it.uniroma2.sag.kelp.predictionfunction

Class SequencePrediction

java.lang.Object

it.uniroma2.sag.kelp.predictionfunction.SequencePrediction

All Implemented Interfaces:

Prediction

public class SequencePrediction
extends Object
implements Prediction

It is a output provided by a machine learning systems on a sequence. This specific implementation allows to assign multiple labelings to single sequence, useful for some labeling strategies, such as Beam Search. Notice that each labeling requires a score to select the more promising labeling.

Author:

Danilo Croce

Constructor Summary

Constructors

Constructor and Description
SequencePrediction()

Method Summary

Methods

Modifier and Type	Method and Description
SequencePath	bestPath()
List<SequencePath>	getPaths()
Float	getScore(Label label) Return the prediction score associated to a given label
void	setPaths(List<SequencePath> paths)
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

SequencePrediction

public SequencePrediction()

Method Detail

bestPath

public SequencePath bestPath()

Returns:

The best path, i.e., the labeling with the highest score in the list of labelings provided by a classifier

getPaths

public List<SequencePath> getPaths()

Returns:

a list containing multiple labelings to be assigned to a single sequence

getScore

public Float getScore(Label label)

Description copied from interface: Prediction

Return the prediction score associated to a given label

Specified by:

getScore in interface Prediction

Parameters:

label - the label whose associated score is required

Returns:

the score associated to a given label

setPaths

public void setPaths(List<SequencePath> paths)

Parameters:

paths - a list contains multiple labelings to be assigned to a single sequence

toString

public String toString()

Overrides:

toString in class Object