Package org.nlpub.watset.graph

Class MarkovClustering<V,E>

java.lang.Object

org.nlpub.watset.graph.MarkovClustering<V,E>

Type Parameters:

V - the type of nodes in the graph

E - the type of edges in the graph

All Implemented Interfaces:

Clustering<V>

public class MarkovClustering<V,E>
extends Object
implements Clustering<V>

Naïve implementation of the Markov Clustering (MCL) algorithm.

This implementation assumes processing of relatively small graphs due to the lack of pruning optimizations.

See Also:

van Dongen (2008)

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
class	MarkovClustering.InflateVisitor	Visitor that raises each element to the power of r.
static class	MarkovClustering.NormalizeVisitor	Visitor that normalizes columns.

Field Summary

Fields

Modifier and Type	Field	Description
protected int	e	The expansion parameter.
protected org.jgrapht.Graph<V,E>	graph	The graph.
protected Map<V,Integer>	index	The mapping of graph nodes to the columns of matrix.
protected MarkovClustering.InflateVisitor	inflateVisitor	The inflation visitor that raises each element of matrix to the power of r.
static Integer	ITERATIONS	The default number of Markov Clustering iterations.
protected org.apache.commons.math3.linear.RealMatrix	matrix	The stochastic matrix.
protected org.apache.commons.math3.linear.RealMatrix	ones	The row matrix filled by ones.
protected double	r	The inflation parameter.

Constructor Summary

Constructors

Constructor	Description
MarkovClustering(org.jgrapht.Graph<V,E> graph, int e, double r)	Create an instance of the Markov Clustering algorithm.

Method Summary

Modifier and Type	Method	Description
protected Map<V,Integer>	buildIndex()	Index the nodes of the input graph.
protected org.apache.commons.math3.linear.RealMatrix	buildMatrix(Map<V,Integer> index)	Construct an adjacency matrix for the given graph.
protected void	expand()	Perform the expansion step.
void	fit()	Run the algorithm to induce the parameters of the clusters.
Collection<Collection<V>>	getClusters()	Return a collection of clusters, each cluster is a collection of objects.
protected void	inflate()	Perform the inflation step.
protected void	normalize()	Normalize the matrix.
static <V,E> Function<org.jgrapht.Graph<V,E>,Clustering<V>>	provider(int e, double r)	A factory function that sets up the algorithm for the given graph.

Methods inherited from class java.lang.Object

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

Field Detail

ITERATIONS

public static final Integer ITERATIONS

The default number of Markov Clustering iterations.

graph

protected final org.jgrapht.Graph<V,E> graph

The graph.

e

protected final int e

The expansion parameter.

r

protected final double r

The inflation parameter.

inflateVisitor

protected final MarkovClustering.InflateVisitor inflateVisitor

The inflation visitor that raises each element of matrix to the power of r.

matrix

protected org.apache.commons.math3.linear.RealMatrix matrix

The stochastic matrix.

ones

protected org.apache.commons.math3.linear.RealMatrix ones

The row matrix filled by ones.

index

protected Map<V,Integer> index

The mapping of graph nodes to the columns of matrix.

Constructor Detail

MarkovClustering

public MarkovClustering(org.jgrapht.Graph<V,E> graph,
                        int e,
                        double r)

Create an instance of the Markov Clustering algorithm.

Parameters:

graph - the graph

e - the expansion parameter

r - the inflation parameter

Method Detail

provider

public static <V,E> Function<org.jgrapht.Graph<V,E>,Clustering<V>> provider(int e,
                                                                                              double r)

A factory function that sets up the algorithm for the given graph.

Type Parameters:

V - the type of nodes in the graph

E - the type of edges in the graph

Parameters:

e - the expansion parameter

r - the inflation parameter

Returns:

a factory function that sets up the algorithm for the given graph

fit

public void fit()

Description copied from interface: Clustering

Run the algorithm to induce the parameters of the clusters.

Specified by:

fit in interface Clustering<V>

getClusters

public Collection<Collection<V>> getClusters()

Description copied from interface: Clustering

Return a collection of clusters, each cluster is a collection of objects. Usually this method is called after the Clustering.fit() method.

Specified by:

getClusters in interface Clustering<V>

Returns:

clusters

buildIndex

protected Map<V,Integer> buildIndex()

Index the nodes of the input graph.

Returns:

a node index

buildMatrix

protected org.apache.commons.math3.linear.RealMatrix buildMatrix(Map<V,Integer> index)

Construct an adjacency matrix for the given graph.

Note that the loops in the graph are ignored.

Parameters:

index - the node index

Returns:

an adjacency matrix

normalize

protected void normalize()

Normalize the matrix.

expand

protected void expand()

Perform the expansion step.

inflate

protected void inflate()

Perform the inflation step.