org.apache.mahout.clustering.cdbw

Class CDbwEvaluator

java.lang.Object

org.apache.mahout.clustering.cdbw.CDbwEvaluator

public final class CDbwEvaluator
extends Object

This class calculates the CDbw metric as defined in http://www.db-net.aueb.gr/index.php/corporate/content/download/227/833/file/HV_poster2002.pdf

Constructor Summary

Constructors

Constructor and Description
CDbwEvaluator(org.apache.hadoop.conf.Configuration conf, org.apache.hadoop.fs.Path clustersIn) Initialize a new instance from job information
CDbwEvaluator(Map<Integer,List<VectorWritable>> representativePoints, List<Cluster> clusters, DistanceMeasure measure) For testing only

Method Summary

Methods

Modifier and Type	Method and Description
double	getCDbw() Compute the CDbw validity metric (eqn 8).
Map<Integer,Map<Integer,Double>>	interClusterDensities() This function evaluates the density of points in the regions between each clusters (eqn 1).
double	interClusterDensity() This function evaluates the average density of points in the regions between clusters (eqn 1).
Vector	intraClusterDensities() The average density within clusters is defined as the percentage of representative points that reside in the neighborhood of the clusters' centers.
double	intraClusterDensity() The average density within clusters is defined as the percentage of representative points that reside in the neighborhood of the clusters' centers.
double	separation() Calculate the separation of clusters (eqn 4) taking into account both the distances between the clusters' closest points and the Inter-cluster density.

Methods inherited from class java.lang.Object

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

Constructor Detail

CDbwEvaluator

public CDbwEvaluator(Map<Integer,List<VectorWritable>> representativePoints,
             List<Cluster> clusters,
             DistanceMeasure measure)

For testing only

Parameters:

representativePoints - a Map> of representative points keyed by clusterId

clusters - a Map of the clusters keyed by clusterId

measure - an appropriate DistanceMeasure

CDbwEvaluator

public CDbwEvaluator(org.apache.hadoop.conf.Configuration conf,
             org.apache.hadoop.fs.Path clustersIn)

Initialize a new instance from job information

Parameters:

conf - a Configuration with appropriate parameters

clustersIn - a String path to the input clusters directory

Method Detail

getCDbw

public double getCDbw()

Compute the CDbw validity metric (eqn 8). The goal of this metric is to reward clusterings which have a high intraClusterDensity and also a high cluster separation.

Returns:

a double

intraClusterDensity

public double intraClusterDensity()

The average density within clusters is defined as the percentage of representative points that reside in the neighborhood of the clusters' centers. The goal is the density within clusters to be significantly high. (eqn 5)

Returns:

a double

interClusterDensities

public Map<Integer,Map<Integer,Double>> interClusterDensities()

This function evaluates the density of points in the regions between each clusters (eqn 1). The goal is the density in the area between clusters to be significant low.

Returns:

a Map> of the inter-cluster densities

separation

public double separation()

Calculate the separation of clusters (eqn 4) taking into account both the distances between the clusters' closest points and the Inter-cluster density. The goal is the distances between clusters to be high while the representative point density in the areas between them are low.

Returns:

a double

interClusterDensity

public double interClusterDensity()

This function evaluates the average density of points in the regions between clusters (eqn 1). The goal is the density in the area between clusters to be significant low.

Returns:

a double

intraClusterDensities

public Vector intraClusterDensities()

The average density within clusters is defined as the percentage of representative points that reside in the neighborhood of the clusters' centers. The goal is the density within clusters to be significantly high. (eqn 5)

Returns:

a Vector of the intra-densities of each clusterId