Package elki.clustering.kmedoids

Class ReynoldsPAM.Instance

java.lang.Object

elki.clustering.kmedoids.PAM.Instance

elki.clustering.kmedoids.ReynoldsPAM.Instance

Enclosing class:

ReynoldsPAM<O>

protected static class ReynoldsPAM.Instance
extends PAM.Instance

Instance for a single dataset.

Author:

Erich Schubert

Field Summary

Fields inherited from class elki.clustering.kmedoids.PAM.Instance

assignment, distQ, ids, nearest, second

Constructor Summary

Constructors

Constructor	Description
Instance(DistanceQuery<?> distQ, DBIDs ids, WritableIntegerDataStore assignment)	Constructor.

Method Summary

Modifier and Type	Method	Description
protected double	computeReassignmentCost(DBIDRef h, WritableDoubleDataStore tnearest)	Compute the reassignment cost, for all medoids in one pass.
protected double	computeRemovalCost(int i, WritableDoubleDataStore tnearest)	Compute the cost of removing a medoid just once.
protected double	run(ArrayModifiableDBIDs medoids, int maxiter)	Run the PAM optimization phase.

Methods inherited from class elki.clustering.kmedoids.PAM.Instance

assignToNearestCluster, computeReassignmentCost

Methods inherited from class java.lang.Object

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

Constructor Detail

Instance

public Instance(DistanceQuery<?> distQ,
                DBIDs ids,
                WritableIntegerDataStore assignment)

Constructor.

Parameters:

distQ - Distance query

ids - IDs to process

assignment - Cluster assignment

Method Detail

run

protected double run(ArrayModifiableDBIDs medoids,
                     int maxiter)

Run the PAM optimization phase.

Overrides:

run in class PAM.Instance

Parameters:

medoids - Medoids list

maxiter -

Returns:

final cost

computeRemovalCost

protected double computeRemovalCost(int i,
                                    WritableDoubleDataStore tnearest)

Compute the cost of removing a medoid just once. This can then be reused for every point, thus decreasing the runtime cost at low memory overhead.

The output array contains for each medoid the cost of removing all its points, and reassigning them to the second nearest medoid instead.

Parameters:

i - Medoid to virtually remove

tnearest - Output storage

Returns:

Cost

computeReassignmentCost

protected double computeReassignmentCost(DBIDRef h,
                                         WritableDoubleDataStore tnearest)

Compute the reassignment cost, for all medoids in one pass.

Parameters:

h - Current object to swap with any medoid.

tnearest - Distance to the nearest except the removed medoid

Returns:

cost change