edu.northwestern.at.wordhoard.swing.calculator.analysis
Class MultiwordUnitData

java.lang.Object
  edu.northwestern.at.wordhoard.swing.calculator.analysis.MultiwordUnitData

public class MultiwordUnitData
extends java.lang.Object

Multiword unit data.

Holds data on counts and association measure values for one multiword unit.

Field Summary

protected double	dice
protected NGramExtractor[]	extractors
protected java.lang.String	leftSuccessorPattern
protected double	logLikelihood
protected java.lang.String	mwu
protected int	mwuCount
protected int	mwuLength
protected double	phiSquared
protected java.lang.String	rightSuccessorPattern
protected double	scp
protected double	si
protected double	sigLogLikelihood
protected int	totalWordCount
protected java.util.Map	wordCountMap
protected int[]	wordCounts
protected java.lang.String[]	words

Constructor Summary

MultiwordUnitData(java.lang.String mwu, java.util.Map wordCountMap, int totalWordCount, NGramExtractor[] extractors)

Method Summary

void	calculateAssociationMeasures() Calculate the association measures.
double	freq(java.lang.String[] words, int i1, int i2) Calculate the frequency for a portion of a set of words.
protected double	getAvp() Get the fair dispersion point normalization.
protected double	getAvp2() Get the fair dispersion point normalization.
double	getAvx() Calculate fair probability for the left hand side of a pseudo-bigram.
double	getAvy() Calculate fair probability for the right hand side of a pseudo-bigram.
double	getDice() Return the Dice coefficient.
double	getLogLikelihood() Return log likelihood.
java.lang.String	getMWUText() Get the multiword unit text.
int	getMWUTextCount() Get the count for this multiword unit text.
int	getMWUTextLength() Get the number of words in this multiword unit.
double	getPhiSquared() Return phi squared.
double	getSCP() Return the symmetric conditional probability.
double	getSI() Return the specific mutual information.
double	getSigLogLikelihood() Return significance of log likelihood.
int	getWordCount(java.lang.String word) Get count for a specific word from the count map.
int[]	getWordCounts() Get the count for each word in this multiword unit.
java.lang.String[]	getWords() Get the words in this multiword unit.
java.lang.String	leftAntecedent() Get the left antecedent of the current multiword unit.
java.lang.String[]	leftSuccessors() Get the left successors of the current multiword unit.
double	prob(java.lang.String[] words, int i1, int i2) Calculate the probability for a portion of a set of words.
java.lang.String	rightAntecedent() Get the right antecedent of the current multiword unit.
java.lang.String[]	rightSuccessors() Get the right successors of the current multiword unit.
java.lang.String[]	successors() Get the successors of the current multiword unit.
java.lang.String	toString() Return mwu as a displayable string.

Methods inherited from class java.lang.Object

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

Field Detail

mwu

protected java.lang.String mwu

mwuCount

protected int mwuCount

mwuLength

protected int mwuLength

words

protected java.lang.String[] words

wordCounts

protected int[] wordCounts

dice

protected double dice

logLikelihood

protected double logLikelihood

phiSquared

protected double phiSquared

scp

protected double scp

si

protected double si

sigLogLikelihood

protected double sigLogLikelihood

extractors

protected NGramExtractor[] extractors

leftSuccessorPattern

protected java.lang.String leftSuccessorPattern

rightSuccessorPattern

protected java.lang.String rightSuccessorPattern

totalWordCount

protected int totalWordCount

wordCountMap

protected java.util.Map wordCountMap

Constructor Detail

MultiwordUnitData

public MultiwordUnitData(java.lang.String mwu,
                         java.util.Map wordCountMap,
                         int totalWordCount,
                         NGramExtractor[] extractors)

Method Detail

getMWUText

public java.lang.String getMWUText()

Get the multiword unit text.

Returns:

The multiword unit text.

getMWUTextCount

public int getMWUTextCount()

Get the count for this multiword unit text.

Returns:

Count of appearances of this multiword unit.

getMWUTextLength

public int getMWUTextLength()

Get the number of words in this multiword unit.

Returns:

Number of words in this multiword unit.

getWords

public java.lang.String[] getWords()

Get the words in this multiword unit.

Returns:

Words in this multiword unit.

getWordCounts

public int[] getWordCounts()

Get the count for each word in this multiword unit.

Returns:

Count for each word in this multiword unit.

leftAntecedent

public java.lang.String leftAntecedent()

Get the left antecedent of the current multiword unit.

Returns:

The left antecedent as a string.

rightAntecedent

public java.lang.String rightAntecedent()

Get the right antecedent of the current multiword unit.

Returns:

The right antecedent as a string.

successors

public java.lang.String[] successors()

Get the successors of the current multiword unit.

Returns:

The successors as an array of strings.

leftSuccessors

public java.lang.String[] leftSuccessors()

Get the left successors of the current multiword unit.

Returns:

The left successors as an array of strings.

rightSuccessors

public java.lang.String[] rightSuccessors()

Get the right successors of the current multiword unit.

Returns:

The right successors as an array of strings.

getAvx

public double getAvx()

Calculate fair probability for the left hand side of a pseudo-bigram.

Returns:

Fair probability for left hand side of pseudo-bigram.

getAvy

public double getAvy()

Calculate fair probability for the right hand side of a pseudo-bigram.

Returns:

Fair probability for right hand side of pseudo-bigram.

getAvp

protected double getAvp()

Get the fair dispersion point normalization.

Returns:

Fair dispersion point normalization.

getAvp2

protected double getAvp2()

Get the fair dispersion point normalization.

Returns:

Fair dispersion point normalization.

calculateAssociationMeasures

public void calculateAssociationMeasures()

Calculate the association measures.

prob

public double prob(java.lang.String[] words,
                   int i1,
                   int i2)

Calculate the probability for a portion of a set of words.

Parameters:

words - The words.

i1 - Starting index.

i2 - Ending index.

Returns:

Probability from word counts.

We use the maximum likelihood estimate of the probability, which is just the number of times the word appears divided by the number of words. For ngrams, we divide the number of times the ngram appears by the total number of ngrams containing the same number of words.

freq

public double freq(java.lang.String[] words,
                   int i1,
                   int i2)

Calculate the frequency for a portion of a set of words.

Parameters:

words - The words.

i1 - Starting index.

i2 - Ending index.

Returns:

Frequency from ngram frequencies.

getDice

public double getDice()

Return the Dice coefficient.

Returns:

The Dice coefficient.

getLogLikelihood

public double getLogLikelihood()

Return log likelihood.

Returns:

log likelihood.

getPhiSquared

public double getPhiSquared()

Return phi squared.

Returns:

phi squared.

getSCP

public double getSCP()

Return the symmetric conditional probability.

Returns:

The symmetric conditional probability.

getSI

public double getSI()

Return the specific mutual information.

Returns:

The specific mutual information.

getSigLogLikelihood

public double getSigLogLikelihood()

Return significance of log likelihood.

Returns:

significance of log likelihood.

getWordCount

public int getWordCount(java.lang.String word)

Get count for a specific word from the count map.

Parameters:

word - The word text.

Returns:

The count for the specified word. 0 if the word does not occur.

toString

public java.lang.String toString()

Return mwu as a displayable string.

Overrides:

toString in class java.lang.Object

Returns:

The mwu as a displayable string.