edu.illinois.cs.cogcomp.ner.ClassifiersAndUtils

Class MemoryEfficientNB

java.lang.Object

edu.illinois.cs.cogcomp.ner.ClassifiersAndUtils.Classifier

edu.illinois.cs.cogcomp.ner.ClassifiersAndUtils.MemoryEfficientNB

public class MemoryEfficientNB
extends Classifier

Field Summary

Fields

Modifier and Type	Field and Description
double[]	classCounts
double	fidCount
Vector<Hashtable<Integer,Double>>	fidCounts
FeatureMap	map
double	sampleSize
static double	smooth
double[]	weights
double[]	wordCounts

Fields inherited from class edu.illinois.cs.cogcomp.ner.ClassifiersAndUtils.Classifier

classesN, methodName

Constructor Summary

Constructors

Constructor and Description
MemoryEfficientNB(DocumentCollection docs, FeatureMap _map, int _classesN)
MemoryEfficientNB(FeatureMap _map, int _classesN)
MemoryEfficientNB(String file)

Method Summary

Modifier and Type	Method and Description
void	allocateSpace(FeatureMap _map, int _classesN)
int	classify(Document doc, double thres)
double	getAcc(DocumentCollection test)
String	getExtendedFeatures(Document d)
double	getFidProb(int fid, int classId)
double[]	getPredictionConfidence(Document doc)
double	getPrior(int classId)
Hashtable<String,Integer>	getTopPmiWords(int maxWordsPerClass, double confThres, int minAppThres)
CharacteristicWords	getTopPmiWords(int classId, int maxWordsPerClass, double confThres, int minAppThres)
void	onlineLearning(Document doc)
void	save(String file)
static boolean	toBeKept(Vector<String> tokens, Hashtable<String,Integer> coolWords, double minRatio, int minLen)
void	weightedOnlineLearning(int[] activeFeatures, double weight, int classID)

Methods inherited from class java.lang.Object

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

Field Detail

smooth

public static final double smooth

See Also:

Constant Field Values

sampleSize

public double sampleSize

weights

public double[] weights

wordCounts

public double[] wordCounts

fidCount

public double fidCount

classCounts

public double[] classCounts

fidCounts

public Vector<Hashtable<Integer,Double>> fidCounts

map

public FeatureMap map

Constructor Detail

MemoryEfficientNB

public MemoryEfficientNB(String file)

MemoryEfficientNB

public MemoryEfficientNB(FeatureMap _map,
                         int _classesN)

MemoryEfficientNB

public MemoryEfficientNB(DocumentCollection docs,
                         FeatureMap _map,
                         int _classesN)

Method Detail

save

public void save(String file)

onlineLearning

public void onlineLearning(Document doc)

weightedOnlineLearning

public void weightedOnlineLearning(int[] activeFeatures,
                                   double weight,
                                   int classID)

classify

public int classify(Document doc,
                    double thres)

Specified by:

classify in class Classifier

getPredictionConfidence

public double[] getPredictionConfidence(Document doc)

Specified by:

getPredictionConfidence in class Classifier

allocateSpace

public void allocateSpace(FeatureMap _map,
                          int _classesN)

getFidProb

public double getFidProb(int fid,
                         int classId)

getPrior

public double getPrior(int classId)

getAcc

public double getAcc(DocumentCollection test)

getTopPmiWords

public Hashtable<String,Integer> getTopPmiWords(int maxWordsPerClass,
                                                double confThres,
                                                int minAppThres)

getTopPmiWords

public CharacteristicWords getTopPmiWords(int classId,
                                          int maxWordsPerClass,
                                          double confThres,
                                          int minAppThres)

toBeKept

public static boolean toBeKept(Vector<String> tokens,
                               Hashtable<String,Integer> coolWords,
                               double minRatio,
                               int minLen)

getExtendedFeatures

public String getExtendedFeatures(Document d)

Specified by:

getExtendedFeatures in class Classifier