SimpleCluster (Stanford JavaNLP API)

Overview

Package

Class

Tree

Deprecated

Index

Help

PREV CLASS NEXT CLASS

FRAMES NO FRAMES

SUMMARY: NESTED | FIELD | CONSTR | METHOD

DETAIL: FIELD | CONSTR | METHOD

edu.stanford.nlp.cluster
Class SimpleCluster

java.lang.Object
  |
  +--edu.stanford.nlp.cluster.SimpleCluster

All Implemented Interfaces:: Cloneable, Cluster

Direct Known Subclasses:: HardCluster, HiddenState, SoftCluster

public class SimpleCluster
extends Object
implements Cluster

A simple implementation of Cluster. Warning: Does not initialize Arrays pr_w_z and pr_d_z upon construction Subclasses should have constructor initialize Arrays pr_w_z and pr_d_z

Field Summary

protected int index
          cluster index

protected Array mean
          weighted mean of datum in cluster

protected int nd
          number of data

protected int nt
          number of features

protected Array pr_d_z
          P(d|z) for a set of w: Probability distribution of data over cluster

protected Array pr_w_z
          P(w|z) for a set of w: Probability distribution of features over cluster

protected double pr_z
          P(z): Probability of given cluster

protected double scatter
          scatter value of data in cluster.

Constructor Summary

SimpleCluster(double prz, Array prwz, Array prdz, int i)


SimpleCluster(int num_features, int num_data)


SimpleCluster(int num_features, int num_data, int i)


Method Summary

void clearData()
          Dets the value for all data in cluster = 0

Object clone()
          Returns deep copy of Cluster

boolean equals(Object o)
          Implements equality test.

double evaluateIntraSimilarity(Matrix m, Array mean)
          Evaluates the cohesiveness of Cluster A low value reflects a cohesive cluster, and a high value reflects a scattered cluster;

Array evaluateWeightedMean(Matrix m)
          Evaluates weighted mean of columns in Matrix m.

Array get_pr_d_z()
          Returns entire probabilitiy distribution P(w|z)

double get_pr_d_z(int datum_index)
          Returns P(d|z) for d=datum_index

Array get_pr_w_z()
          Returns entire probability distribution P(w|z)

double get_pr_w_z(int feature_index)
          Returns P(w|z) for w=feature_index

double get_pr_z()
          Returns P(z)

int getIndex()
          Returns cluster's index

double getIntraSimilarity()
          Returns scatter value of Cluster, if it has already been computed

Array getMean()
          Returns weighted mean of Cluster, if it has already been computed

void set_pr_d_z(Array prdz)
          Sets entire P(d|z) array to prdz

void set_pr_d_z(int datum_index, double value)
          Sets P(d|z) for d=datum_index to value P(d|z) is the probability of datum d, given this cluster

void set_pr_w_z(Array prwz)
          Sets entire P(w|z) array to prwz

void set_pr_w_z(int feature_index, double value)
          Sets P(w|z) for w=feature_index to value P(w|z) is the probability of feature w, given this cluster

void set_pr_z(double value)
          Sets P(z) to value P(z) is the cluster probability, or cluster weight

void setIndex(int i)
          Sets cluster index to i

String toString(DataCollection dbm)
          Returns String representation of Cluster prints default number of features and data.

String toString(int tnf, int tnd, DataCollection dbm)
          Returns String representation of Cluster

String toXMLString(DataCollection dbm)
          Returns XML String representation of Cluster prints default number of features and data.

String toXMLString(int tnf, int tnd, DataCollection dbm)
          Returns XML String representation of Cluster

Methods inherited from class java.lang.Object

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

Field Detail

index

protected int index

cluster index

scatter

protected double scatter

scatter value of data in cluster. if data in cluster are very scattered, scatter will be high.

mean

protected Array mean

weighted mean of datum in cluster

pr_z

protected double pr_z

P(z): Probability of given cluster

pr_w_z

protected Array pr_w_z

P(w|z) for a set of w: Probability distribution of features over cluster

pr_d_z

protected Array pr_d_z

P(d|z) for a set of w: Probability distribution of data over cluster

nt

protected int nt

number of features

nd

protected int nd

number of data

Constructor Detail

SimpleCluster

public SimpleCluster(int num_features,
                     int num_data)

SimpleCluster

public SimpleCluster(int num_features,
                     int num_data,
                     int i)

SimpleCluster

public SimpleCluster(double prz,
                     Array prwz,
                     Array prdz,
                     int i)

Method Detail