io.kgraph.library.clustering

Class SemiClustering

java.lang.Object

io.kgraph.library.clustering.SemiClustering

All Implemented Interfaces:

ComputeFunction<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>>

public class SemiClustering
extends Object
implements ComputeFunction<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>>

Implements the Semi-Clustering algorithm as presented in the Pregel paper from SIGMOD'10.

The input to the algorithm is an undirected weighted graph and the output is a set of clusters with each vertex potentially belonging to multiple clusters.

A semi-cluster is assigned a score S=(I-f*B)/(V(V-1)/2), where I is the sum of weights of all internal edges, B is the sum of weights of all boundary edges, V is the number of vertices in the semi-cluster, f is a user-specified boundary edge score factor with a value between 0 and 1.

Each vertex maintains a list containing a maximum number of semi-clusters, sorted by score. The lists gets greedily updated in an iterative manner.

The algorithm finishes when the semi-cluster lists don't change or after a maximum number of iterations.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SemiClustering.SemiCluster This class represents a semi-cluster.

Nested classes/interfaces inherited from interface io.kgraph.pregel.ComputeFunction

ComputeFunction.Aggregators, ComputeFunction.Callback<K,VV,EV,Message>, ComputeFunction.InitCallback, ComputeFunction.MasterCallback, ComputeFunction.ReadAggregators, ComputeFunction.ReadWriteAggregators

Field Summary

Fields

Modifier and Type	Field and Description
static String	CLUSTER_CAPACITY Maximum number of vertices in a semi-cluster.
static int	CLUSTER_CAPACITY_DEFAULT Default value for cluster capacity.
static String	ITERATIONS Maximum number of iterations.
static int	ITERATIONS_DEFAULT Default value for ITERATIONS.
static String	MAX_CLUSTERS Maximum number of semi-clusters.
static int	MAX_CLUSTERS_DEFAULT Default value for maximum number of semi-clusters.
static String	SCORE_FACTOR Boundary edge score factor.
static double	SCORE_FACTOR_DEFAULT Default value for Boundary Edge Score Factor.

Constructor Summary

Constructors

Constructor and Description
SemiClustering()

Method Summary

Modifier and Type	Method and Description
void	compute(int superstep, VertexWithValue<Long,Set<SemiClustering.SemiCluster>> vertex, Iterable<Set<SemiClustering.SemiCluster>> messages, Iterable<EdgeWithValue<Long,Double>> edges, ComputeFunction.Callback<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>> cb) Compute method.
void	init(Map<String,?> configs, ComputeFunction.InitCallback cb) Initialize the ComputeFunction, this is the place to register aggregators.

Methods inherited from class java.lang.Object

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

Methods inherited from interface io.kgraph.pregel.ComputeFunction

masterCompute, postSuperstep, preSuperstep

Field Detail

ITERATIONS

public static final String ITERATIONS

Maximum number of iterations.

See Also:

Constant Field Values

ITERATIONS_DEFAULT

public static final int ITERATIONS_DEFAULT

Default value for ITERATIONS.

See Also:

Constant Field Values

MAX_CLUSTERS

public static final String MAX_CLUSTERS

Maximum number of semi-clusters.

See Also:

Constant Field Values

MAX_CLUSTERS_DEFAULT

public static final int MAX_CLUSTERS_DEFAULT

Default value for maximum number of semi-clusters.

See Also:

Constant Field Values

CLUSTER_CAPACITY

public static final String CLUSTER_CAPACITY

Maximum number of vertices in a semi-cluster.

See Also:

Constant Field Values

CLUSTER_CAPACITY_DEFAULT

public static final int CLUSTER_CAPACITY_DEFAULT

Default value for cluster capacity.

See Also:

Constant Field Values

SCORE_FACTOR

public static final String SCORE_FACTOR

Boundary edge score factor.

See Also:

Constant Field Values

SCORE_FACTOR_DEFAULT

public static final double SCORE_FACTOR_DEFAULT

Default value for Boundary Edge Score Factor.

See Also:

Constant Field Values

Constructor Detail

SemiClustering

public SemiClustering()

Method Detail

init

public final void init(Map<String,?> configs,
                       ComputeFunction.InitCallback cb)

Description copied from interface: ComputeFunction

Initialize the ComputeFunction, this is the place to register aggregators.

Specified by:

init in interface ComputeFunction<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>>

Parameters:

configs - configuration parameters

cb - a callback for registering aggregators

compute

public void compute(int superstep,
                    VertexWithValue<Long,Set<SemiClustering.SemiCluster>> vertex,
                    Iterable<Set<SemiClustering.SemiCluster>> messages,
                    Iterable<EdgeWithValue<Long,Double>> edges,
                    ComputeFunction.Callback<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>> cb)

Compute method.

Specified by:

compute in interface ComputeFunction<Long,Set<SemiClustering.SemiCluster>,Double,Set<SemiClustering.SemiCluster>>

Parameters:

messages - Messages received

superstep - the count of the current superstep

vertex - the current vertex with its value

edges - the adjacent edges with their values

cb - a callback for setting a new vertex value or sending messages to the next superstep