org.dice_research.topicmodeling.io

Class ZipfBasedTopWordWriter

java.lang.Object

org.dice_research.topicmodeling.io.AbstractModelFilesWriter

org.dice_research.topicmodeling.io.ZipfBasedTopWordWriter

All Implemented Interfaces:

CSVFileProcessor, ModelWriter

public class ZipfBasedTopWordWriter
extends AbstractModelFilesWriter
implements CSVFileProcessor

Field Summary

Fields

Modifier and Type	Field and Description
private String	fileNamePrefix
private static double	FREQUENCY_CLASS_WEIGHTING
private static double	LOG2
private static org.slf4j.Logger	LOGGER
private static int	MAX_NUMBER_OF_TOP_WORDS
private org.dice_research.topicmodeling.algorithms.ProbTopicModelingAlgorithmStateSupplier	probTopicStateSupplier
private static String	SEPARATING_TOP_WORDS_FILE

Fields inherited from class org.dice_research.topicmodeling.io.AbstractModelFilesWriter

folder

Fields inherited from interface org.dice_research.topicmodeling.io.CSVFileProcessor

ESCAPECHAR, QUOTECHAR, SEPARATOR

Constructor Summary

Constructors

Constructor and Description
ZipfBasedTopWordWriter(File folder, String fileNamePrefix)
ZipfBasedTopWordWriter(org.dice_research.topicmodeling.algorithms.ProbTopicModelingAlgorithmStateSupplier probTopicStateSupplier, File folder)

Method Summary

Modifier and Type	Method and Description
private int[]	determineBestWords(int[] wordCounts, com.carrotsearch.hppc.IntDoubleOpenHashMap frequencyClasses, com.carrotsearch.hppc.IntDoubleOpenHashMap differences, double maxBackgroundClass, double maxForegroundClass)
private double	determineFrequencyClasses(int[] wordCounts, com.carrotsearch.hppc.IntDoubleOpenHashMap frequencyClasses)
protected com.carrotsearch.hppc.IntDoubleOpenHashMap	getClassDifferences(com.carrotsearch.hppc.IntDoubleOpenHashMap foregroundFrequencyClasses, com.carrotsearch.hppc.IntDoubleOpenHashMap backgroundFrequencyClasses, double maxBackgroundClass)
private int[]	getCountDifference(int[] backgroundCount, int[] foregroundCount)
protected int[]	getFacetsForSingleTopic(int[] foregroundWordCounts, int[] backgroundWordCounts)
private double	getMaxCount(int[] wordCounts)
void	writeModelToFiles(org.dice_research.topicmodeling.algorithms.Model model)
private void	writeTopWords(int[][] topWords, org.dice_research.topicmodeling.utils.vocabulary.Vocabulary vocabulary)

Methods inherited from class org.dice_research.topicmodeling.io.AbstractModelFilesWriter

getFolder, setFolder

Methods inherited from class java.lang.Object

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

Field Detail

LOGGER

private static final org.slf4j.Logger LOGGER

SEPARATING_TOP_WORDS_FILE

private static final String SEPARATING_TOP_WORDS_FILE

See Also:

Constant Field Values

MAX_NUMBER_OF_TOP_WORDS

private static final int MAX_NUMBER_OF_TOP_WORDS

See Also:

Constant Field Values

FREQUENCY_CLASS_WEIGHTING

private static final double FREQUENCY_CLASS_WEIGHTING

See Also:

Constant Field Values

LOG2

private static final double LOG2

fileNamePrefix

private String fileNamePrefix

probTopicStateSupplier

private org.dice_research.topicmodeling.algorithms.ProbTopicModelingAlgorithmStateSupplier probTopicStateSupplier

Constructor Detail

ZipfBasedTopWordWriter

public ZipfBasedTopWordWriter(org.dice_research.topicmodeling.algorithms.ProbTopicModelingAlgorithmStateSupplier probTopicStateSupplier,
                              File folder)

ZipfBasedTopWordWriter

public ZipfBasedTopWordWriter(File folder,
                              String fileNamePrefix)

Method Detail

writeModelToFiles

public void writeModelToFiles(org.dice_research.topicmodeling.algorithms.Model model)

Specified by:

writeModelToFiles in interface ModelWriter

writeTopWords

private void writeTopWords(int[][] topWords,
                           org.dice_research.topicmodeling.utils.vocabulary.Vocabulary vocabulary)

getFacetsForSingleTopic

protected int[] getFacetsForSingleTopic(int[] foregroundWordCounts,
                                        int[] backgroundWordCounts)

determineFrequencyClasses

private double determineFrequencyClasses(int[] wordCounts,
                                         com.carrotsearch.hppc.IntDoubleOpenHashMap frequencyClasses)

getCountDifference

private int[] getCountDifference(int[] backgroundCount,
                                 int[] foregroundCount)

getMaxCount

private double getMaxCount(int[] wordCounts)

getClassDifferences

protected com.carrotsearch.hppc.IntDoubleOpenHashMap getClassDifferences(com.carrotsearch.hppc.IntDoubleOpenHashMap foregroundFrequencyClasses,
                                                                         com.carrotsearch.hppc.IntDoubleOpenHashMap backgroundFrequencyClasses,
                                                                         double maxBackgroundClass)

determineBestWords

private int[] determineBestWords(int[] wordCounts,
                                 com.carrotsearch.hppc.IntDoubleOpenHashMap frequencyClasses,
                                 com.carrotsearch.hppc.IntDoubleOpenHashMap differences,
                                 double maxBackgroundClass,
                                 double maxForegroundClass)