org.tribuo.multilabel.evaluation

Class MultiLabelConfusionMatrix

java.lang.Object

org.tribuo.multilabel.evaluation.MultiLabelConfusionMatrix

All Implemented Interfaces:

ConfusionMatrix<MultiLabel>

public final class MultiLabelConfusionMatrix
extends Object
implements ConfusionMatrix<MultiLabel>

A ConfusionMatrix which accepts MultiLabels.

In a multi-label confusion matrix M,

 tn = M[:, 0, 0]
 fn = M[:, 0, 1]
 fp = M[:, 1, 0]
 tp = M[:, 1, 1]
 

For class-wise values,

 tn(class i) = M[i, 0, 0]
 fn(class i) = M[i, 0, 1]
 fp(class i) = M[i, 1, 0]
 tp(class i) = M[i, 1, 1]
 

Constructor Summary

Constructors

Constructor and Description
MultiLabelConfusionMatrix(Model<MultiLabel> model, List<Prediction<MultiLabel>> predictions)

Method Summary

Modifier and Type	Method and Description
double	confusion(MultiLabel predicted, MultiLabel truth) The number of times the supplied predicted label was returned for the supplied true class.
double	fn(MultiLabel cls) The number of false negatives for the supplied label.
double	fp(MultiLabel cls) The number of false positives for the supplied label.
ImmutableOutputInfo<MultiLabel>	getDomain() Returns the classification domain that this confusion matrix operates over.
double	support() The number of examples this confusion matrix has seen.
double	support(MultiLabel cls) The number of examples with this true label this confusion matrix has seen.
double	tn(MultiLabel cls) The number of true negatives for the supplied label.
String	toString()
double	tp(MultiLabel cls) The number of true positives for the supplied label.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.tribuo.classification.evaluation.ConfusionMatrix

fn, fp, sumOverOutputs, tn, tp

Constructor Detail

MultiLabelConfusionMatrix

public MultiLabelConfusionMatrix(Model<MultiLabel> model,
                                 List<Prediction<MultiLabel>> predictions)

Method Detail

support

public double support(MultiLabel cls)

Description copied from interface: ConfusionMatrix

The number of examples with this true label this confusion matrix has seen.

Specified by:

support in interface ConfusionMatrix<MultiLabel>

Parameters:

cls - The label.

Returns:

The number of examples.

getDomain

public ImmutableOutputInfo<MultiLabel> getDomain()

Description copied from interface: ConfusionMatrix

Returns the classification domain that this confusion matrix operates over.

Specified by:

getDomain in interface ConfusionMatrix<MultiLabel>

Returns:

The classification domain.

support

public double support()

Description copied from interface: ConfusionMatrix

The number of examples this confusion matrix has seen.

Specified by:

support in interface ConfusionMatrix<MultiLabel>

Returns:

The number of examples.

tp

public double tp(MultiLabel cls)

Description copied from interface: ConfusionMatrix

The number of true positives for the supplied label.

Specified by:

tp in interface ConfusionMatrix<MultiLabel>

Parameters:

cls - The label.

Returns:

The number of examples.

fp

public double fp(MultiLabel cls)

Description copied from interface: ConfusionMatrix

The number of false positives for the supplied label.

Specified by:

fp in interface ConfusionMatrix<MultiLabel>

Parameters:

cls - The label.

Returns:

The number of examples.

fn

public double fn(MultiLabel cls)

Description copied from interface: ConfusionMatrix

The number of false negatives for the supplied label.

Specified by:

fn in interface ConfusionMatrix<MultiLabel>

Parameters:

cls - The label.

Returns:

The number of examples.

tn

public double tn(MultiLabel cls)

Description copied from interface: ConfusionMatrix

The number of true negatives for the supplied label.

Specified by:

tn in interface ConfusionMatrix<MultiLabel>

Parameters:

cls - The label.

Returns:

The number of examples.

confusion

public double confusion(MultiLabel predicted,
                        MultiLabel truth)

Description copied from interface: ConfusionMatrix

The number of times the supplied predicted label was returned for the supplied true class.

Specified by:

confusion in interface ConfusionMatrix<MultiLabel>

Parameters:

predicted - The predicted label.

truth - The true label.

Returns:

The number of examples predicted as predictedLabel when the true label was trueLabel.

toString

public String toString()

Overrides:

toString in class Object