Package edu.illinois.cs.cogcomp.edison.features.factory

Class Summary

Class	Description
BrownClusterFeatureExtractor	A WordFeatureExtractor that generates prefixes of brown cluster Ids for each word.
ChunkEmbedding	Adds the following features using chunks from the specified view: The number of chunks contained in the constituent A thresholded version of the number of chunks.
ChunkPathPattern	Extracts the following features from two constituents The concatenation of shallow parse chunks between the last token of the constituent to the left and the first token of the other one.
ChunkPropertyFeatureFactory	A collection of features that extract properties from chunks.
ClauseFeatureExtractor
CorelexFeatureExtractor	Returns a set of features that are CoreLex's basic types of lemmas.
DependencyPath	Given two tokens, this feature extractor extracts the following features: The path (without tokens) in the dependency tree from the first token of the first constituent to the first token of the second one. The path, where tokens are represented by their POS tag The length of the path
DependencyPathNgrams	Given two tokens, this feature extractor extracts the following features: The path in the dependency tree from the first token of the first constituent to the first token of the second one. The length of the path
LevinVerbClassFeature	Adds a collection of features that correspond to Levin's verb classes, as defined in "English Verb Classes And Alternations: A Preliminary Investigation"
LinearDistance	Given two constituents, indicates distance between the two.
LinearPosition	Given two constituents, indicates whether the first is before or after the other or have an overlap.
ListFeatureFactory
NomLexClassFeature	Adds NomLex based features.
ParseHeadWordPOS	Extracts the head word and POS tag of the constituent.
ParseLabelIdentifier	This abstract feature extractor checks if the constituent's label in the parse tree satisfies some property.
ParsePath	Given two constituents, this feature extractor extracts the following features: The path in the parse tree from the one constituent to another The length of the parse path
ParsePhraseType	Extracts the label of the phrase and that of its parent along with the head word of the parent according to the Collins' head percolation rules.
ParsePhraseTypeOnly	Extracts only the label of the phrase.
ParseSiblings	Extracts the phrase type, head word and head POS of the left and right siblings of the input constituent.
RogetThesaurusFeatures	A BIG dictionary of words (and multi-word phrases) and their synonyms.
SpanFeaturesOrdered
SpanFeaturesUnordered
SpanLengthFeature
SubcategorizationFrame	Given a constituent, this feature finds the corresponding node in the parse tree and reports expands its parent.
SyntacticFrame	This feature is described in Xue and Palmer 2004.
VerbVoiceIndicator	Uses the parse tree to figure out whether a verb is an active or a passive voice.
WordFeatureExtractorFactory
WordNetFeatureExtractor	This feature extractor generates the following WordNet based features from a word: synonyms, synsets, hypernyms, hypernym-sets.

Enum Summary

Enum	Description
WordNetFeatureExtractor.WordNetFeatureClass	Feature types as used in WordNetFeatureExtractor.addFeatureType(WordNetFeatureClass).