com.wcohen.ss.abbvGapsHmm

Class AlignmentPredictionModel

java.lang.Object

com.wcohen.ss.abbvGapsHmm.AlignmentPredictionModel

public class AlignmentPredictionModel
extends Object

Author:

Dana Movshovitz-Attias

Field Summary

Fields

Modifier and Type	Field and Description
static String	_trainingCorpusFile
static String	_trainingDataDir
static String	_trueLabelsFile
static String	SEPARATOR

Constructor Summary

Constructors

Constructor and Description
AlignmentPredictionModel()

Method Summary

Modifier and Type	Method and Description
Map<String,Acronym>	acronymsArrayToMap(Collection<Acronym> pairs)
protected void	addCandidatePair(List<Acronym> allPairs, String longFormCandidate, String shortFormCandidate)
protected String	chunkLongForm(String longForm, int size)
List<Acronym>	extractCandidatePairs(String text) Candidates Extraction
protected List<Acronym>	extractHeadNounPattern_2Parts(String text)
protected List<Acronym>	extractHeadNounPattern_3Parts(String text)
protected List<Acronym>	extractPatternAcronyms(String text)
protected List<Acronym>	extractSingleAcronyms(String text)
protected List<Acronym>	extractTrailingNounPattern_2Parts(String text)
protected List<Acronym>	extractTrailingNounPattern_3Parts(String text)
List<Double>	getEmmisions()
List<Double>	getTransitions()
protected boolean	isShortForm(String candidate)
protected boolean	isValidExpression(String exp)
protected boolean	isValidShortForm(String candidate)
static ArrayList<Map<String,String>>	loadLabels(String labelsFile)
static List<String>	loadTrainingCorpus(String corpusFile)
protected Acronym	parseCandidate(String outOfParenthesis, String inParenthesis)
Acronym	predict(Acronym candidatePair)
Collection<Acronym>	predict(String text)
AbbreviationAlignmentContainer<AbbvGapsHMM.Emissions,AbbvGapsHMM.States>	predict(String sf, String lf)
AbbreviationAlignmentContainer<AbbvGapsHMM.Emissions,AbbvGapsHMM.States>	predictAlignment(Acronym candidatePair)
void	setModelParamsFile()
void	setModelParamsFile(String paramFilename)
void	setStartingParams(List<Double> emmisions, List<Double> transitions)
void	setTfIdfData(String dataFile)
void	setTrainingDataDir(String trainDir)
boolean	train(List<String> corpus, List<Integer> trainingSet, List<Map<String,String>> trueLabels)
boolean	trainIfNeeded()
boolean	trainOnAll()
boolean	trainOnCandidates()

Methods inherited from class java.lang.Object

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

Field Detail

SEPARATOR

public static final String SEPARATOR

See Also:

Constant Field Values

_trainingDataDir

public static String _trainingDataDir

_trueLabelsFile

public static String _trueLabelsFile

_trainingCorpusFile

public static String _trainingCorpusFile

Constructor Detail

AlignmentPredictionModel

public AlignmentPredictionModel()
                         throws IOException

Throws:

IOException

Method Detail

setTrainingDataDir

public void setTrainingDataDir(String trainDir)

setTfIdfData

public void setTfIdfData(String dataFile)
                  throws IOException

Throws:

IOException

setModelParamsFile

public void setModelParamsFile(String paramFilename)

setModelParamsFile

public void setModelParamsFile()

loadLabels

public static ArrayList<Map<String,String>> loadLabels(String labelsFile)

loadTrainingCorpus

public static List<String> loadTrainingCorpus(String corpusFile)

trainOnAll

public boolean trainOnAll()

trainOnCandidates

public boolean trainOnCandidates()

train

public boolean train(List<String> corpus,
                     List<Integer> trainingSet,
                     List<Map<String,String>> trueLabels)

predict

public AbbreviationAlignmentContainer<AbbvGapsHMM.Emissions,AbbvGapsHMM.States> predict(String sf,
                                                                                        String lf)

predictAlignment

public AbbreviationAlignmentContainer<AbbvGapsHMM.Emissions,AbbvGapsHMM.States> predictAlignment(Acronym candidatePair)

predict

public Acronym predict(Acronym candidatePair)

acronymsArrayToMap

public Map<String,Acronym> acronymsArrayToMap(Collection<Acronym> pairs)

predict

public Collection<Acronym> predict(String text)

trainIfNeeded

public boolean trainIfNeeded()

extractCandidatePairs

public List<Acronym> extractCandidatePairs(String text)

Candidates Extraction

extractPatternAcronyms

protected List<Acronym> extractPatternAcronyms(String text)

addCandidatePair

protected void addCandidatePair(List<Acronym> allPairs,
                                String longFormCandidate,
                                String shortFormCandidate)

extractHeadNounPattern_3Parts

protected List<Acronym> extractHeadNounPattern_3Parts(String text)

extractHeadNounPattern_2Parts

protected List<Acronym> extractHeadNounPattern_2Parts(String text)

extractTrailingNounPattern_3Parts

protected List<Acronym> extractTrailingNounPattern_3Parts(String text)

extractTrailingNounPattern_2Parts

protected List<Acronym> extractTrailingNounPattern_2Parts(String text)

extractSingleAcronyms

protected List<Acronym> extractSingleAcronyms(String text)

parseCandidate

protected Acronym parseCandidate(String outOfParenthesis,
                                 String inParenthesis)

chunkLongForm

protected String chunkLongForm(String longForm,
                               int size)

isValidExpression

protected boolean isValidExpression(String exp)

isShortForm

protected boolean isShortForm(String candidate)

isValidShortForm

protected boolean isValidShortForm(String candidate)

getEmmisions

public List<Double> getEmmisions()

getTransitions

public List<Double> getTransitions()

setStartingParams

public void setStartingParams(List<Double> emmisions,
                              List<Double> transitions)