edu.illinois.cs.cogcomp.edison.features

Class WordFeatureExtractor

java.lang.Object

edu.illinois.cs.cogcomp.edison.features.WordFeatureExtractor

All Implemented Interfaces:

FeatureExtractor

Direct Known Subclasses:

BrownClusterFeatureExtractor, CorelexFeatureExtractor, NgramFeatureExtractor, PrefixFeatureExtractor, SuffixFeatureExtractor, VerbVoiceIndicator, WordNetFeatureExtractor

public abstract class WordFeatureExtractor
extends Object
implements FeatureExtractor

Features of a word.

If the input constituent is not a word, then the feature extractor can do one of two things: If a flag is set in the constructor, then it will generate features from the last word of the constituent. If the flag is not set, then it will throw a EdisonException.

Author:

Vivek Srikumar

Constructor Summary

Constructors

Constructor and Description
WordFeatureExtractor() Creates a new WordFeatureExtractor.
WordFeatureExtractor(boolean useLastWordOfMultiwordConstituents) Creates a new WordFeatureExtractor.

Method Summary

Modifier and Type	Method and Description
static WordFeatureExtractor	convertToWordFeatureExtractor(FeatureExtractor fex)
Set<Feature>	getFeatures(Constituent c)
String	getName()
abstract Set<Feature>	getWordFeatures(TextAnnotation ta, int wordPosition)

Methods inherited from class java.lang.Object

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

Constructor Detail

WordFeatureExtractor

public WordFeatureExtractor(boolean useLastWordOfMultiwordConstituents)

Creates a new WordFeatureExtractor. If parameter useLastWordOfMultiwordConstituents is true, then the feature extractor will generate features from the last word of multi-word constituents. If it is not true, then the feature extractor will throw an exception on seeing a multi-word constituent.

It is probably safest to the parameter to true. This will provide a check to ensure that the WordFeatureExtractor only sees words.

WordFeatureExtractor

public WordFeatureExtractor()

Creates a new WordFeatureExtractor. This constructor is equivalent to calling new WordFeatureExtractor(true).

See Also:

WordFeatureExtractor(boolean)

Method Detail

convertToWordFeatureExtractor

public static WordFeatureExtractor convertToWordFeatureExtractor(FeatureExtractor fex)

getFeatures

public Set<Feature> getFeatures(Constituent c)
                         throws EdisonException

Specified by:

getFeatures in interface FeatureExtractor

Throws:

EdisonException

getWordFeatures

public abstract Set<Feature> getWordFeatures(TextAnnotation ta,
                                             int wordPosition)
                                      throws EdisonException

Throws:

EdisonException

getName

public String getName()

Specified by:

getName in interface FeatureExtractor