org.openimaj.experiment.evaluation.classification

Class BasicClassificationResult<CLASS>

java.lang.Object

org.openimaj.experiment.evaluation.classification.BasicClassificationResult<CLASS>

Type Parameters:

CLASS - type of class predicted by the Classifier

All Implemented Interfaces:

ClassificationResult<CLASS>

public class BasicClassificationResult<CLASS>
extends Object
implements ClassificationResult<CLASS>

A basic implementation of a ClassificationResult that internally maintains a map of classes to confidences.

A threshold is used to determine whether a class has a high-enough confidence to be considered part of the predicted set of classes.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Constructor Summary

Constructors

Constructor and Description
BasicClassificationResult() Construct with a default threshold of 0.
BasicClassificationResult(double threshold) Construct with the given threshold.

Method Summary

Modifier and Type	Method and Description
double	getConfidence(CLASS clazz) Get the confidence associated with the given class.
Set<CLASS>	getPredictedClasses() Get the set of classes predicted by this result.
void	put(CLASS clz, double confidence) Add a class/confidence pair.

Methods inherited from class java.lang.Object

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

Constructor Detail

BasicClassificationResult

public BasicClassificationResult()

Construct with a default threshold of 0.

BasicClassificationResult

public BasicClassificationResult(double threshold)

Construct with the given threshold.

Parameters:

threshold - the threshold

Method Detail

put

public void put(CLASS clz,
                double confidence)

Add a class/confidence pair.

Parameters:

clz - the class

confidence - the confidence

getConfidence

public double getConfidence(CLASS clazz)

Description copied from interface: ClassificationResult

Get the confidence associated with the given class. If the class is unknown, then this method should return 0.

Normally, confidence values are bounded between 0 and 1.

Specified by:

getConfidence in interface ClassificationResult<CLASS>

Parameters:

clazz - the target class.

Returns:

the confidence in assigning the class.

getPredictedClasses

public Set<CLASS> getPredictedClasses()

Description copied from interface: ClassificationResult

Get the set of classes predicted by this result. Usually, this is a subset of all classes; specifically those with a high confidence.

Specified by:

getPredictedClasses in interface ClassificationResult<CLASS>

Returns:

the set of predicted classes.