de.lmu.ifi.dbs.elki.algorithm.clustering.correlation
Class COPAC<V extends NumberVector<V,?>,D extends Distance<D>>

java.lang.Object
  de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm<Clustering<Model>>
      de.lmu.ifi.dbs.elki.algorithm.clustering.correlation.COPAC<V,D>

Type Parameters:

V - the type of NumberVector handled by this Algorithm

All Implemented Interfaces:

Algorithm, ClusteringAlgorithm<Clustering<Model>>, InspectionUtilFrequentlyScanned, Parameterizable

@Title(value="COPAC: COrrelation PArtition Clustering")
@Description(value="Partitions a database according to the correlation dimension of its objects and performs a clustering algorithm over the partitions.")
@Reference(authors="E. Achtert, C. B\u00f6hm, H.-P. Kriegel, P. Kr\u00f6ger P., A. Zimek",
           title="Robust, Complete, and Efficient Correlation Clustering",
           booktitle="Proc. 7th SIAM International Conference on Data Mining (SDM\'07), Minneapolis, MN, 2007",
           url="http://www.siam.org/proceedings/datamining/2007/dm07_037achtert.pdf")
public class COPAC<V extends NumberVector<V,?>,D extends Distance<D>>
extends AbstractAlgorithm<Clustering<Model>>
implements ClusteringAlgorithm<Clustering<Model>>

Provides the COPAC algorithm, an algorithm to partition a database according to the correlation dimension of its objects and to then perform an arbitrary clustering algorithm over the partitions.

Reference: Achtert E., Böhm C., Kriegel H.-P., Kröger P., Zimek A.: Robust, Complete, and Efficient Correlation Clustering.
In Proc. 7th SIAM International Conference on Data Mining (SDM'07), Minneapolis, MN, 2007

Nested Class Summary

static class

COPAC.Parameterizer<V extends NumberVector<V,?>,D extends Distance<D>> Parameterization class.

Field Summary

private static Logging	logger The logger for this class.
static OptionID	PARTITION_ALGORITHM_ID Parameter to specify the clustering algorithm to apply to each partition, must extend ClusteringAlgorithm.
static OptionID	PARTITION_DISTANCE_ID Parameter to specify the distance function to use inside the partitions AbstractIndexBasedDistanceFunction .
private Class<? extends ClusteringAlgorithm<Clustering<Model>>>	partitionAlgorithm Get the algorithm to run on each partition.
private Collection<Pair<OptionID,Object>>	partitionAlgorithmParameters Holds the parameters of the algorithm to run on each partition.
private FilteredLocalPCABasedDistanceFunction<V,?,D>	partitionDistanceFunction Holds the instance of the preprocessed distance function PARTITION_DISTANCE_ID.
private FilteredLocalPCABasedDistanceFunction.Instance<V,LocalProjectionIndex<V,?>,D>	partitionDistanceQuery The last used distance query
static OptionID	PREPROCESSOR_ID Parameter to specify the local PCA preprocessor to derive partition criterion, must extend AbstractFilteredPCAIndex.

Constructor Summary

COPAC(FilteredLocalPCABasedDistanceFunction<V,?,D> partitionDistanceFunction, Class<? extends ClusteringAlgorithm<Clustering<Model>>> partitionAlgorithm, Collection<Pair<OptionID,Object>> partitionAlgorithmParameters)
Constructor.

Method Summary

TypeInformation[]	getInputTypeRestriction() Get the input type restriction used for negotiating the data query.
protected Logging	getLogger() Get the (STATIC) logger for this class.
ClusteringAlgorithm<Clustering<Model>>	getPartitionAlgorithm(DistanceQuery<V,D> query) Returns the partition algorithm.
FilteredLocalPCABasedDistanceFunction.Instance<V,LocalProjectionIndex<V,?>,D>	getPartitionDistanceQuery() Get the last used distance query (to expose access to the preprocessor) Used by ERiC.
Clustering<Model>	run(Relation<V> relation) Performs the COPAC algorithm on the given database.
private Clustering<Model>	runPartitionAlgorithm(Relation<V> relation, Map<Integer,DBIDs> partitionMap, DistanceQuery<V,D> query) Runs the partition algorithm and creates the result.

Methods inherited from class de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm

makeParameterDistanceFunction, run

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.lmu.ifi.dbs.elki.algorithm.clustering.ClusteringAlgorithm

run

Field Detail

logger

private static final Logging logger

The logger for this class.

PREPROCESSOR_ID

public static final OptionID PREPROCESSOR_ID

Parameter to specify the local PCA preprocessor to derive partition criterion, must extend AbstractFilteredPCAIndex.

Key: -copac.preprocessor

PARTITION_DISTANCE_ID

public static final OptionID PARTITION_DISTANCE_ID

Parameter to specify the distance function to use inside the partitions AbstractIndexBasedDistanceFunction .

Default value: LocallyWeightedDistanceFunction

Key: -copac.partitionDistance

PARTITION_ALGORITHM_ID

public static final OptionID PARTITION_ALGORITHM_ID

Parameter to specify the clustering algorithm to apply to each partition, must extend ClusteringAlgorithm.

Key: -copac.partitionAlgorithm

partitionDistanceFunction

private FilteredLocalPCABasedDistanceFunction<V extends NumberVector<V,?>,?,D extends Distance<D>> partitionDistanceFunction

Holds the instance of the preprocessed distance function PARTITION_DISTANCE_ID.

partitionAlgorithm

private Class<? extends ClusteringAlgorithm<Clustering<Model>>> partitionAlgorithm

Get the algorithm to run on each partition.

partitionAlgorithmParameters

private Collection<Pair<OptionID,Object>> partitionAlgorithmParameters

Holds the parameters of the algorithm to run on each partition.

partitionDistanceQuery

private FilteredLocalPCABasedDistanceFunction.Instance<V extends NumberVector<V,?>,LocalProjectionIndex<V extends NumberVector<V,?>,?>,D extends Distance<D>> partitionDistanceQuery

The last used distance query

Constructor Detail

COPAC

public COPAC(FilteredLocalPCABasedDistanceFunction<V,?,D> partitionDistanceFunction,
             Class<? extends ClusteringAlgorithm<Clustering<Model>>> partitionAlgorithm,
             Collection<Pair<OptionID,Object>> partitionAlgorithmParameters)

Constructor.

Parameters:

partitionDistanceFunction - Distance function

partitionAlgorithm - Algorithm to use on partitions

partitionAlgorithmParameters - Parameters for Algorithm to run on partitions

Method Detail

run

public Clustering<Model> run(Relation<V> relation)
                      throws IllegalStateException

Performs the COPAC algorithm on the given database.

Parameters:

relation - Relation to process

Returns:

Clustering result

Throws:

IllegalStateException

runPartitionAlgorithm

private Clustering<Model> runPartitionAlgorithm(Relation<V> relation,
                                                Map<Integer,DBIDs> partitionMap,
                                                DistanceQuery<V,D> query)

Runs the partition algorithm and creates the result.

Parameters:

relation - the database to run this algorithm on

partitionMap - the map of partition IDs to object ids

query - The preprocessor based query function

getPartitionAlgorithm

public ClusteringAlgorithm<Clustering<Model>> getPartitionAlgorithm(DistanceQuery<V,D> query)

Returns the partition algorithm.

Returns:

the specified partition algorithm

getPartitionDistanceQuery

public FilteredLocalPCABasedDistanceFunction.Instance<V,LocalProjectionIndex<V,?>,D> getPartitionDistanceQuery()

Get the last used distance query (to expose access to the preprocessor) Used by ERiC. TODO: migrate to factory pattern!

Returns:

distance query

getInputTypeRestriction

public TypeInformation[] getInputTypeRestriction()

Description copied from class: AbstractAlgorithm

Get the input type restriction used for negotiating the data query.

Specified by:

getInputTypeRestriction in interface Algorithm

Specified by:

getInputTypeRestriction in class AbstractAlgorithm<Clustering<Model>>

Returns:

Type restriction

getLogger

protected Logging getLogger()

Description copied from class: AbstractAlgorithm

Get the (STATIC) logger for this class.

Specified by:

getLogger in class AbstractAlgorithm<Clustering<Model>>

Returns:

the static logger