de.lmu.ifi.dbs.elki.evaluation.clustering

Class Entropy

java.lang.Object

de.lmu.ifi.dbs.elki.evaluation.clustering.Entropy

@Reference(authors="M. Meil\u0103",
           title="Comparing clusterings by the variation of information",
           booktitle="Learning theory and kernel machines",
           url="https://doi.org/10.1007/978-3-540-45167-9_14",
           bibkey="DBLP:conf/colt/Meila03")
public class Entropy
extends java.lang.Object

Entropy based measures.

References:

M. Meilă
Comparing clusterings by the variation of information
Learning theory and kernel machines

Since:

0.5.0

Author:

Sascha Goldhofer

Field Summary

Fields

Modifier and Type	Field and Description
protected double	entropyFirst Entropy in first
protected double	entropyJoint Joint entropy
protected double	entropySecond Entropy in second

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	Entropy(ClusterContingencyTable table) Constructor.

Method Summary

Modifier and Type	Method and Description
double	entropyConditionalFirst() Get the conditional entropy of the first clustering.
double	entropyConditionalSecond() Get the conditional entropy of the first clustering.
double	entropyFirst() Get the entropy of the first clustering using Log_2.
double	entropyJoint() Get the joint entropy of both clusterings (not normalized, 0 = equal)
double	entropyMutualInformation() Get the mutual information (not normalized, 0 = equal)
double	entropyNMIJoint() Get the joint-normalized mutual information (normalized, 0 = unequal)
double	entropyNMIMax() Get the max-normalized mutual information (normalized, 0 = unequal)
double	entropyNMIMin() Get the min-normalized mutual information (normalized, 0 = unequal)
double	entropyNMISqrt() Get the sqrt-normalized mutual information (normalized, 0 = unequal)
double	entropyNMISum() Get the sum-normalized mutual information (normalized, 0 = unequal)
double	entropyPowers() Get Powers entropy (normalized, 0 = equal) Powers = 1 - NMI_Sum
double	entropySecond() Get the entropy of the second clustering using Log_2.
double	normalizedVariationOfInformation() Get the normalized variation of information (normalized, 0 = equal) NVI = 1 - NMI_Joint X.
double	variationOfInformation() Get the variation of information (not normalized, 0 = equal)

Methods inherited from class java.lang.Object

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

Field Detail

entropyFirst

protected double entropyFirst

Entropy in first

entropySecond

protected double entropySecond

Entropy in second

entropyJoint

protected double entropyJoint

Joint entropy

Constructor Detail

Entropy

protected Entropy(ClusterContingencyTable table)

Constructor.

Parameters:

table - Contingency table

Method Detail

entropyFirst

public double entropyFirst()

Get the entropy of the first clustering using Log_2. (not normalized, 0 = equal)

Returns:

Entropy of first clustering

entropySecond

public double entropySecond()

Get the entropy of the second clustering using Log_2. (not normalized, 0 = equal)

Returns:

Entropy of second clustering

entropyJoint

public double entropyJoint()

Get the joint entropy of both clusterings (not normalized, 0 = equal)

Returns:

Joint entropy of both clusterings

entropyConditionalFirst

public double entropyConditionalFirst()

Get the conditional entropy of the first clustering. (not normalized, 0 = equal)

Returns:

Conditional entropy of first clustering

entropyConditionalSecond

public double entropyConditionalSecond()

Get the conditional entropy of the first clustering. (not normalized, 0 = equal)

Returns:

Conditional entropy of second clustering

entropyPowers

public double entropyPowers()

Get Powers entropy (normalized, 0 = equal) Powers = 1 - NMI_Sum

Returns:

Powers

entropyMutualInformation

public double entropyMutualInformation()

Get the mutual information (not normalized, 0 = equal)

Returns:

Mutual information

entropyNMIJoint

public double entropyNMIJoint()

Get the joint-normalized mutual information (normalized, 0 = unequal)

Returns:

Joint Normalized Mutual information

entropyNMIMin

public double entropyNMIMin()

Get the min-normalized mutual information (normalized, 0 = unequal)

Returns:

Min Normalized Mutual information

entropyNMIMax

public double entropyNMIMax()

Get the max-normalized mutual information (normalized, 0 = unequal)

Returns:

Max Normalized Mutual information

entropyNMISum

public double entropyNMISum()

Get the sum-normalized mutual information (normalized, 0 = unequal)

Returns:

Sum Normalized Mutual information

entropyNMISqrt

public double entropyNMISqrt()

Get the sqrt-normalized mutual information (normalized, 0 = unequal)

Returns:

Sqrt Normalized Mutual information

variationOfInformation

public double variationOfInformation()

Get the variation of information (not normalized, 0 = equal)

Returns:

Variation of information

normalizedVariationOfInformation

@Reference(authors="X. V. Nguyen, J. Epps, J. Bailey",
           title="Information theoretic measures for clusterings comparison: is a correction for chance necessary?",
           booktitle="Proc. 26th Ann. Int. Conf. on Machine Learning (ICML \'09)",
           url="https://doi.org/10.1145/1553374.1553511",
           bibkey="DBLP:conf/icml/NguyenEB09")
public double normalizedVariationOfInformation()

Get the normalized variation of information (normalized, 0 = equal) NVI = 1 - NMI_Joint

X. V. Nguyen, J. Epps, J. Bailey
Information theoretic measures for clusterings comparison: is a correction for chance necessary?
Proc. 26th Ann. Int. Conf. on Machine Learning (ICML '09)

Returns:

Normalized Variation of information