CommunityComparisons (Cognitive Foundry)

java.lang.Object
- gov.sandia.cognition.graph.community.CommunityComparisons<NodeNameType>

Type Parameters:: NodeNameType - The type used to name nodes

@PublicationReference(author="Nguyen Xuan Vinh, Julien Epps, and James Bailey",
                      title="Information Theoretic Measures of Clusterings Comparison: Variants, Properties, Normalization and Correction for Chance",
                      type=Journal,
                      year=2010,
                      publication="Journal of Machine Learning Research",
                      pages=-17)
public class CommunityComparisons<NodeNameType>
extends java.lang.Object

This class implements various comparisons of partitionings of a graph (or the resulting communities)

Author:: jdwendt

Constructor Summary

Constructors
Constructor and Description
`CommunityComparisons(NodePartitioning<NodeNameType> u, NodePartitioning<NodeNameType> v)` Initializes this with the input partitionings of the same underlying set

Method Summary

All Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`double`	`getAmiMax()` Returns the max adjusted-for-chance mutual information (see Table 2 of cited paper)
`double`	`getAmiMin()` Returns the min adjusted-for-chance mutual information (see Table 2 of cited paper)
`double`	`getAmiSqrt()` Returns the sqrt adjusted-for-chance mutual information (see Table 2 of cited paper)
`double`	`getAmiSum()` Returns the sum adjusted-for-chance mutual information (see Table 2 of cited paper)
`double`	`getConditionalEntropyUGivenV()` Returns H(u\|v), computing and storing it if this is the first time called
`double`	`getConditionalEntropyVGivenU()` Returns H(v\|u), computing and storing it if this is the first time called
`double`	`getEntropyU()` Returns the entropy of the u partitioning (computing and storing it if this is the first time called)
`double`	`getEntropyV()` Returns the entropy of the v partitioning (computing and storing it if this is the first time called)
`double`	`getExpectedMutualInformation()` Returns the approximate expected value of the mutual information for the two partitionings.
`double`	`getJointEntropy()` Returns the joint entropy (H(u,v)), computing it if this is the first time called
`double`	`getMutualInformation()` Returns the mutual information (I(u, v)) for the two partitionings, computing and storing it if this is the first time called
`double`	`getNmiJoint()` Returns the joint normalized mutual information (see Table 2 of cited paper)
`double`	`getNmiMax()` Returns the max normalized mutual information (see Table 2 of cited paper)
`double`	`getNmiMin()` Returns the min normalized mutual information (see Table 2 of cited paper)
`double`	`getNmiSqrt()` Returns the sqrt normalized mutual information (see Table 2 of cited paper)
`double`	`getNmiSum()` Returns the sum normalized mutual information (see Table 2 of cited paper)

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Constructor Detail
  - CommunityComparisons
```
public CommunityComparisons(NodePartitioning<NodeNameType> u,
                            NodePartitioning<NodeNameType> v)
```
    Initializes this with the input partitionings of the same underlying set
    
    Parameters:
    
    u - The first partitioning
    
    v - The second partitioning
- Method Detail
  - getEntropyU
```
public double getEntropyU()
```
    Returns the entropy of the u partitioning (computing and storing it if this is the first time called)
    
    Returns:
    
    the entropy of the u partitioning
  - getEntropyV
```
public double getEntropyV()
```
    Returns the entropy of the v partitioning (computing and storing it if this is the first time called)
    
    Returns:
    
    the entropy of the v partitioning
  - getJointEntropy
```
public double getJointEntropy()
```
    Returns the joint entropy (H(u,v)), computing it if this is the first time called
    
    Returns:
    
    the joint entropy
  - getConditionalEntropyUGivenV
```
public double getConditionalEntropyUGivenV()
```
    Returns H(u|v), computing and storing it if this is the first time called
    
    Returns:
    
    H(u|v)
  - getConditionalEntropyVGivenU
```
public double getConditionalEntropyVGivenU()
```
    Returns H(v|u), computing and storing it if this is the first time called
    
    Returns:
    
    H(v|u)
  - getMutualInformation
```
public double getMutualInformation()
```
    Returns the mutual information (I(u, v)) for the two partitionings, computing and storing it if this is the first time called
    
    Returns:
    
    the mutual information
  - getNmiJoint
```
public double getNmiJoint()
```
    Returns the joint normalized mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the joint normalized mutual information
  - getNmiMax
```
public double getNmiMax()
```
    Returns the max normalized mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the max normalized mutual information
  - getNmiSum
```
public double getNmiSum()
```
    Returns the sum normalized mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the sum normalized mutual information
  - getNmiSqrt
```
public double getNmiSqrt()
```
    Returns the sqrt normalized mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the sqrt normalized mutual information
  - getNmiMin
```
public double getNmiMin()
```
    Returns the min normalized mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the min normalized mutual information
  - getExpectedMutualInformation
```
public double getExpectedMutualInformation()
```
    Returns the approximate expected value of the mutual information for the two partitionings. Specifically, the middle value from Equation 4 of the cited paper. Computes and stores the value if this is the first time called.
    
    Returns:
    
    the approximate expected value of the mutual information
  - getAmiMax
```
public double getAmiMax()
```
    Returns the max adjusted-for-chance mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the max adjusted-for-chance mutual information
  - getAmiSum
```
public double getAmiSum()
```
    Returns the sum adjusted-for-chance mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the sum adjusted-for-chance mutual information
  - getAmiSqrt
```
public double getAmiSqrt()
```
    Returns the sqrt adjusted-for-chance mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the sqrt adjusted-for-chance mutual information
  - getAmiMin
```
public double getAmiMin()
```
    Returns the min adjusted-for-chance mutual information (see Table 2 of cited paper)
    
    Returns:
    
    the min adjusted-for-chance mutual information

Class CommunityComparisons<NodeNameType>

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Constructor Detail

CommunityComparisons

Method Detail

getEntropyU

getEntropyV

getJointEntropy

getConditionalEntropyUGivenV

getConditionalEntropyVGivenU

getMutualInformation

getNmiJoint

getNmiMax

getNmiSum

getNmiSqrt

getNmiMin

getExpectedMutualInformation

getAmiMax

getAmiSum

getAmiSqrt

getAmiMin