javax.datamining.clustering
Interface ClusteringSettings

All Superinterfaces:

BuildSettings, MiningObject

public interface ClusteringSettings

extends BuildSettings

A ClusteringSettings instance holds metadata for required settings common to all clustering algorithms. If maximum number of clusters or minimum cluster case count is not specified, a system determined value is used.

Author:

JSR-73 Java Data Mining Expert Group

See Also:

MiningObject, BuildSettings

Method Summary

AggregationFunction	getAggregationFunction() Returns the aggregation function to be used.
AttributeComparisonFunction	getAttributeComparisonFunction(java.lang.String logicalAttributeName) Returns the attribute comparison function to be used.
long	getMaxClusterCaseCount() Returns the maximum number of cases per cluster for the model to be built.
int	getMaxLevels() Returns the maximum number of levels in a hierarchical model.
int	getMaxNumberOfClusters() Returns the maximum number of clusters in the model to be built.
long	getMinClusterCaseCount() Returns the minimum number of cases per cluster for the model to be built.
SimilarityMatrix	getSimilarityMatrix(java.lang.String logicalAttributeName) Returns the similarity matrix to be used when the attribute comparison function is similarityMatrix.
void	setAggregationFunction(AggregationFunction function) Sets the aggregation function to be used.
void	setAttributeComparisonFunction(java.lang.String logicalAttributeName, AttributeComparisonFunction function) Sets the attribute comparison function to be used.
void	setMaxClusterCaseCount(long maxCount) Sets the maximum number of cases per cluster in the model to be built.
void	setMaxLevels(int numberOfLevels) Sets the maximum level for hierarchical clustering.
void	setMaxNumberOfClusters(int maxClusters) Sets the maximum number of clusters in the model to be built.
void	setMinClusterCaseCount(long minCaseCount) Sets the minimum number of cases per cluster in the model to be built.
void	setSimilarityMatrix(java.lang.String logicalAttributeName, SimilarityMatrix matrix) Sets the similarity matrix to be used.

Methods inherited from interface javax.datamining.base.BuildSettings

getAlgorithmSettings, getDesiredExecutionTimeInMinutes, getLogicalAttributes, getLogicalData, getLogicalDataName, getMiningFunction, getOutlierIdentification, getOutlierTreatment, getUsage, getWeight, getWeightAttribute, setAlgorithmSettings, setDesiredExecutionTimeInMinutes, setLogicalDataName, setOutlierIdentification, setOutlierTreatment, setUsage, setWeight, setWeightAttribute, verify

Methods inherited from interface javax.datamining.MiningObject

getCreationDate, getCreatorInfo, getDescription, getName, getObjectIdentifier, getObjectType, setDescription

Method Detail

getAggregationFunction

public AggregationFunction getAggregationFunction()

Returns the aggregation function to be used.

Returns:

AggregationFunction

getAttributeComparisonFunction

public AttributeComparisonFunction getAttributeComparisonFunction(java.lang.String logicalAttributeName)
                                                           throws JDMException

Returns the attribute comparison function to be used.

Parameters:

logicalAttributeName -

Returns:

AttributeComparisonFunction

Throws:

JDMException

getMaxClusterCaseCount

public long getMaxClusterCaseCount()

Returns the maximum number of cases per cluster for the model to be built.

Returns:

long

getMaxLevels

public int getMaxLevels()

Returns the maximum number of levels in a hierarchical model. Non-hierarchical models have level set to 1. Level must be between 1 and the maximum number of clusters.

Returns:

int

getMaxNumberOfClusters

public int getMaxNumberOfClusters()

Returns the maximum number of clusters in the model to be built.

Returns:

int

getMinClusterCaseCount

public long getMinClusterCaseCount()

Returns the minimum number of cases per cluster for the model to be built.

Returns:

long

getSimilarityMatrix

public SimilarityMatrix getSimilarityMatrix(java.lang.String logicalAttributeName)
                                     throws JDMException

Returns the similarity matrix to be used when the attribute comparison function is similarityMatrix.

Parameters:

logicalAttributeName -

Returns:

SimilarityMatrix

Throws:

JDMException

setAggregationFunction

public void setAggregationFunction(AggregationFunction function)

Sets the aggregation function to be used. If null is specified, the default aggregation function is used.

Parameters:

function - The aggregation function to be used.

Returns:

void

setAttributeComparisonFunction

public void setAttributeComparisonFunction(java.lang.String logicalAttributeName,
                                           AttributeComparisonFunction function)
                                    throws JDMException

Sets the attribute comparison function to be used. Throws an exception on invalid combination of attribute type and comparison function. Valid combinations include absDiff, gaussSim for numerical attributes, delta, equal for categorical attributes, and table for categorical or discrete numerical attributes.

Parameters:

logicalAttributeName -

function - The comparison function to be used for the specified attribute.

Returns:

void

Throws:

JDMException

setMaxClusterCaseCount

public void setMaxClusterCaseCount(long maxCount)

Sets the maximum number of cases per cluster in the model to be built. Combined with minimum cluster case count, this feature is used to control the size of the clusters in the model. The value must be a positive integer.

Parameters:

maxCount - The maximum case count for clusters.

Returns:

void

setMaxLevels

public void setMaxLevels(int numberOfLevels)

Sets the maximum level for hierarchical clustering. The level must be a positive integer.

Parameters:

numberOfLevels - The maximum level allowed in the model.

Returns:

void

setMaxNumberOfClusters

public void setMaxNumberOfClusters(int maxClusters)

Sets the maximum number of clusters in the model to be built. The value must be a positive integer.

Parameters:

maxClusters - The maximum number of clusters to be found during model building.

Returns:

void

setMinClusterCaseCount

public void setMinClusterCaseCount(long minCaseCount)

Sets the minimum number of cases per cluster in the model to be built. If the value is 0, it means there is no minimum cluster count. Otherwise, the value must be a positive integer.

Parameters:

minCaseCount - The minimum cluster size in terms of case count. Any cluster must represent at least this many cases.

Returns:

void

setSimilarityMatrix

public void setSimilarityMatrix(java.lang.String logicalAttributeName,
                                SimilarityMatrix matrix)
                         throws JDMException

Sets the similarity matrix to be used. This method automatically changes the attribute comparison function to similarityMatrix.

Parameters:

logicalAttributeName -

matrix - The similarity matrix to be used for the specified attribute.

Returns:

void

Throws:

JDMException