org.cogcomp.md

Class BIOFeatureExtractor

java.lang.Object

org.cogcomp.md.BIOFeatureExtractor

public class BIOFeatureExtractor
extends Object

This is the FeatureExtractor Class for the lbj file We define most of the features here Each static function performs an extraction on a given target

Constructor Summary

Constructors

Constructor and Description
BIOFeatureExtractor()

Method Summary

Modifier and Type	Method and Description
static List<Pair<Integer,String>>	getBrownClusterPaths(Constituent c) Get the brown-clusters representations of a given Constituent This function checks one word before and two words after it.
static List<String>	getBrownClusterPathsSingle(Constituent c) Extracts brown-cluster representation.
static List<Pair<Integer,String>>	getGazetteerFeatures(Constituent c) Extracts the gazetteers feature for a given constituent The size of the Constituent should be 1.
static List<Pair<String,String>>	getGazetteerFeaturesHead(Relation r) This functions works for the extent classifier.
static List<String>	getGazetteerFeaturesSingle(Constituent c) Extract the gazetteer feature of only the Constituent given.
static List<Pair<Integer,String>>	getPOSFeatures(Constituent c) Extracts Part-of-speech tagging features of a given Constituent The size of the Constituent should be 1
static List<Pair<Integer,String>>	getWordFormFeatures(Constituent c) Get the form/lemma features of a given Constituent.
static String	getWordNetHyms(WordNetManager wordnet, Constituent c) Extracts the hypernyms from wordNet of the given Constituent.
static String	getWordNetTags(WordNetManager wordnet, Constituent c) Extracts the wordNet document names of the given Constituent.
static List<Pair<String,Boolean>>	getWordTypeInformation(Constituent c) This extracts the special form of a given Constituent (all digits...
static String	isInPronounList(Constituent c) Check if the the given Constituent is a pronoun that is defined in a pronoun list Results show that if the list aligns with the definition of the corpus, it helps The size of the Constituent should be 1
static String	isSentenceStart(Constituent c) Check if the Constituent is a start of a sentence The size of the Constituent should be 1

Methods inherited from class java.lang.Object

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

Constructor Detail

BIOFeatureExtractor

public BIOFeatureExtractor()

Method Detail

getGazetteerFeatures

public static List<Pair<Integer,String>> getGazetteerFeatures(Constituent c)

Extracts the gazetteers feature for a given constituent The size of the Constituent should be 1. This function checks the word and two words after it The feature itself was extracted in BIOReader prior to this extraction

getGazetteerFeaturesSingle

public static List<String> getGazetteerFeaturesSingle(Constituent c)

Extract the gazetteer feature of only the Constituent given. No other words is looked at The size of the Constituent should be 1

getGazetteerFeaturesHead

public static List<Pair<String,String>> getGazetteerFeaturesHead(Relation r)

This functions works for the extent classifier. It extracts the gazetteers feature of the head in a given (token,head) pair.

isSentenceStart

public static String isSentenceStart(Constituent c)

Check if the Constituent is a start of a sentence The size of the Constituent should be 1

getWordFormFeatures

public static List<Pair<Integer,String>> getWordFormFeatures(Constituent c)

Get the form/lemma features of a given Constituent. The size of the Constituent should be 1

getPOSFeatures

public static List<Pair<Integer,String>> getPOSFeatures(Constituent c)

Extracts Part-of-speech tagging features of a given Constituent The size of the Constituent should be 1

getWordTypeInformation

public static List<Pair<String,Boolean>> getWordTypeInformation(Constituent c)

This extracts the special form of a given Constituent (all digits... etc) The size of the Constituent should be 1

getBrownClusterPaths

public static List<Pair<Integer,String>> getBrownClusterPaths(Constituent c)

Get the brown-clusters representations of a given Constituent This function checks one word before and two words after it. The size of the Constituent should be 1

getBrownClusterPathsSingle

public static List<String> getBrownClusterPathsSingle(Constituent c)

Extracts brown-cluster representation. It only extracts the feature on the given Constituent The size of the Constituent should be 1

isInPronounList

public static String isInPronounList(Constituent c)

Check if the the given Constituent is a pronoun that is defined in a pronoun list Results show that if the list aligns with the definition of the corpus, it helps The size of the Constituent should be 1

getWordNetTags

public static String getWordNetTags(WordNetManager wordnet,
                                    Constituent c)

Extracts the wordNet document names of the given Constituent. Helps nominals when crossing domains. The size of the Constituent should be 1

getWordNetHyms

public static String getWordNetHyms(WordNetManager wordnet,
                                    Constituent c)

Extracts the hypernyms from wordNet of the given Constituent. Helps nominals when crossing domains. The size of the Constituent should be 1