edu.stanford.nlp.cluster
Class HACM

java.lang.Object
  |
  +--edu.stanford.nlp.cluster.AbstractClusteringMethod
        |
        +--edu.stanford.nlp.cluster.HACM

All Implemented Interfaces:

ClusteringMethod

public class HACM

extends AbstractClusteringMethod

Hierarchical Agglomerative Clustering Method

Field Summary

Fields inherited from class edu.stanford.nlp.cluster.AbstractClusteringMethod

clusters, db, method, nc, nd, nt

Constructor Summary

HACM(Double4Function type)
Sets values for db, nt, nd, csf.

HACM(cern.colt.function.DoubleDoubleFunction type)
Sets values for db, nt, nd, csf.

Method Summary

Entry	closest(Matrix m) Iterates through the similarity matrix to find the clusters with highest similarity.
SimpleClusters	cluster(DataCollection data, int num_clusters) Clusters documents into the desired number of clusters by Hierarchical Agglomerative Clustering
Clusters	cluster(int num_clusters, Double4Function complicatedcsf) Clusters using complicated cluster similarity function, such as group-average
Clusters	cluster(int num_clusters, cern.colt.function.DoubleDoubleFunction simplecsf) Clusters using a simple cluster similarity function, such as single-link or complete link
void	finish() Finalizes clusters by setting P(w|z) to be the weighted mean of the data vectors in the cluster

Methods inherited from class edu.stanford.nlp.cluster.AbstractClusteringMethod

cluster, evaluate, evaluate, initialize, toString, toXMLString

Methods inherited from class java.lang.Object

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

Constructor Detail

HACM

public HACM(cern.colt.function.DoubleDoubleFunction type)

Sets values for db, nt, nd, csf.

Parameters:

type - Cluster Similarity Function (i.e. single-link, complete-link, group average

HACM

public HACM(Double4Function type)

Sets values for db, nt, nd, csf.

Parameters:

type - Cluster Similarity Function (i.e. single-link, complete-link, group average

Method Detail

closest

public Entry closest(Matrix m)

Iterates through the similarity matrix to find the clusters with highest similarity. returns an Entry (i,j,value), where i,j are the two closest clusters, and value is the value assigned by the similarity metric.

cluster

public SimpleClusters cluster(DataCollection data,
                              int num_clusters)

Clusters documents into the desired number of clusters by Hierarchical Agglomerative Clustering

Parameters:

num_clusters - number of final desired clusters

cluster

public Clusters cluster(int num_clusters,
                        cern.colt.function.DoubleDoubleFunction simplecsf)

Clusters using a simple cluster similarity function, such as single-link or complete link

cluster

public Clusters cluster(int num_clusters,
                        Double4Function complicatedcsf)

Clusters using complicated cluster similarity function, such as group-average

finish

public void finish()

Finalizes clusters by setting P(w|z) to be the weighted mean of the data vectors in the cluster