eus.ixa.ixa.pipe.ml.nerc

Class DictionariesNERTagger

java.lang.Object

eus.ixa.ixa.pipe.ml.nerc.DictionariesNERTagger

public class DictionariesNERTagger
extends Object

Named Entity Recognition module based on Dictionaries objects This class provides the following functionalities:

1. string matching against of a string (typically tokens) against a Dictionary containing names. This function is also used to implement Dictionary based features in the training package.
2. tag: Provided a Dictionaries it tags only the names it matches against it
3. post: This function checks for names in the Dictionary that have not been detected by a StatisticalSequenceLabeler; it also corrects the Name type for those detected by a StatisticalSequenceLabeler but also present in a dictionary.

Author:

ragerri 2014/03/14

Constructor Summary

Constructors

Constructor and Description
DictionariesNERTagger(Dictionaries aDictionaries) Construct a DictionaryNameFinder using one dictionary and one named entity class.
DictionariesNERTagger(Dictionaries aDictionaries, SequenceLabelFactory aNameFactory) Construct a DictionariesNameFinder with a dictionary, a type and a name factory.

Method Summary

Modifier and Type	Method and Description
void	clearAdaptiveData() Clear the adaptiveData for each document.
List<SequenceLabel>	getNames(String[] tokens) Dictionaries based Named Entity Detection and Classification.
List<SequenceLabel>	getNamesFromSpans(Span[] neSpans, String[] tokens) Creates a list of SequenceLabel objects from spans and tokens.
Span[]	nercToSpans(String[] tokens) Detects Named Entities in a Dictionaries by NE type ignoring case.
Span[]	nercToSpansExact(String[] tokens) Detects Named Entities in a Dictionaries by NE type This method is case sensitive.

Methods inherited from class java.lang.Object

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

Constructor Detail

DictionariesNERTagger

public DictionariesNERTagger(Dictionaries aDictionaries)

Construct a DictionaryNameFinder using one dictionary and one named entity class.

Parameters:

aDictionaries - the dictionaries

DictionariesNERTagger

public DictionariesNERTagger(Dictionaries aDictionaries,
                             SequenceLabelFactory aNameFactory)

Construct a DictionariesNameFinder with a dictionary, a type and a name factory.

Parameters:

aDictionaries - the dictionaries

aNameFactory - the factory

Method Detail

getNames

public final List<SequenceLabel> getNames(String[] tokens)

Dictionaries based Named Entity Detection and Classification.

Parameters:

tokens - the tokenized sentence

Returns:

a list of detected SequenceLabel objects

nercToSpans

public final Span[] nercToSpans(String[] tokens)

Detects Named Entities in a Dictionaries by NE type ignoring case.

Parameters:

tokens - the tokenized sentence

Returns:

spans of the Named Entities

nercToSpansExact

public final Span[] nercToSpansExact(String[] tokens)

Detects Named Entities in a Dictionaries by NE type This method is case sensitive.

Parameters:

tokens - the tokenized sentence

Returns:

spans of the Named Entities all

getNamesFromSpans

public final List<SequenceLabel> getNamesFromSpans(Span[] neSpans,
                                                   String[] tokens)

Creates a list of SequenceLabel objects from spans and tokens.

Parameters:

neSpans - the spans of the entities in the sentence

tokens - the tokenized sentence

Returns:

a list of SequenceLabel objects

clearAdaptiveData

public void clearAdaptiveData()

Clear the adaptiveData for each document.