com.planetj.taste.impl.recommender
Class NearestNeighborClusterSimilarity

java.lang.Object
  com.planetj.taste.impl.recommender.NearestNeighborClusterSimilarity

All Implemented Interfaces:

Refreshable, ClusterSimilarity

public final class NearestNeighborClusterSimilarity
extends java.lang.Object
implements ClusterSimilarity

Defines cluster similarity as the largest correlation between any two Users in the clusters -- that is, it says that clusters are close when some pair of their members has high correlation.

Since:

1.1

Author:

Sean Owen

Constructor Summary

NearestNeighborClusterSimilarity(UserCorrelation correlation)
Constructs a NearestNeighborClusterSimilarity based on the given UserCorrelation.

NearestNeighborClusterSimilarity(UserCorrelation correlation, double samplingPercentage)
Constructs a NearestNeighborClusterSimilarity based on the given UserCorrelation.

Method Summary

double	getSimilarity(java.util.Collection<User> cluster1, java.util.Collection<User> cluster2)
void	refresh() Triggers "refresh" -- whatever that means -- of the implementation.
java.lang.String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

NearestNeighborClusterSimilarity

public NearestNeighborClusterSimilarity(UserCorrelation correlation)

Constructs a NearestNeighborClusterSimilarity based on the given UserCorrelation. All user-user correlations are examined.

Parameters:

correlation -

NearestNeighborClusterSimilarity

public NearestNeighborClusterSimilarity(UserCorrelation correlation,
                                        double samplingPercentage)

Constructs a NearestNeighborClusterSimilarity based on the given UserCorrelation. By setting samplingPercentage to a value less than 1.0, this implementation will only examine that fraction of all user-user correlations between two clusters, increasing performance at the expense of accuracy.

Parameters:

correlation -

samplingPercentage -

Since:

1.4.5

Method Detail

getSimilarity

public double getSimilarity(java.util.Collection<User> cluster1,
                            java.util.Collection<User> cluster2)
                     throws TasteException

Specified by:

getSimilarity in interface ClusterSimilarity

Parameters:

cluster1 - first cluster of Users

cluster2 - second cluste rof Users

Returns:

"distance" between clusters; a positiv

Throws:

TasteException - if an error occurs while computing similarity, such as errors accessing an underlying DataModel

refresh

public void refresh()

Triggers "refresh" -- whatever that means -- of the implementation. The general contract is that any Refreshable should always leave itself in a consistent, operational state, and that the refresh atomically updates internal state from old to new.

Specified by:

refresh in interface Refreshable

toString

@NotNull
public java.lang.String toString()

Overrides:

toString in class java.lang.Object