Package edu.mit.jmwe.detect

Provides MWE detector API, a baseline detector, plus numerous other detector implementations.

See:
          Description

Interface Summary
IHasMWEDetector A MWE Detector that wraps another MWE detector.
IInflectionRule Represents a pattern according to which the parts of a multi-word expression may be inflected.
IMWEDetector A detector for multi-word expressions.
IMWEDetectorFilter An interface for MWE detectors that act as filters for other MWE detectors
IMWEDetectorResolver An interface for MWE detectors that resolve conflicts between identified MWEs from other detectors.
 

Class Summary
Baseline A baseline implementation of the IMWEDetector interface.
CompositeDetector A detector composed of multiple simpler detectors.
Consecutive Detector that uses an index to detect the MWEs whose parts appear continuously in the sentence.
ConstrainLength A detector filter that constrains MWEs by length.
Continuous A detector that filters the results of its backing detector to return only those MWEs whose parts are continuous in the sentence.
Exhaustive Implements an exhaustive algorithm that detects all possible non-stop-word MWEs in a sentence, including MWEs that are out of order or discontinuous.
FilterByScore Removes all MWEs from the wrapped detector's results whose score is "no good", where "no good" is implementation dependent.
HasMWEDetector Abstract base class for MWE Detectors that wrap, and constraint the output of, other MWE detectors.
InflectionLookup Uses a given MWE detector to find multi-word expressions in a sentence but discards inflected multi-word expressions whose form is not listed as a valid inflected form by its associated IMWEDesc.
InflectionPattern Uses a given IMWEDetector to find multi-word expressions in a sentence but discards inflected multi-word expressions if the inflected form does not correspond to an inflection pattern give by a set of inflection rules.
InOrder Filters the output of a given MWE detector, removing those MWEs whose parts appear do not appear in the sentence in the same order as they appear in the MWE description.
Leftmost A resolver that chooses the leftmost MWE.
LeskAtLeast A filter detector that throws out MWEs whose lesk score is less than some value.
LMLR Stands for "Longest Match, Left to Right".
Longest A detector resolver that chooses the longest MWEs in a set of conflicts.
MoreFrequentAsMWE Filters the results of its backing detector to include only those MWEs whose parts, in a continuous run of tokens, are more often marked as an MWE than as individual tokens.
MWEBuilder<T extends IToken> A record that is used to hold tokens as the detector passes over a sentence.
NoInflection A detector that filters the results of its backing detector to return only those MWEs that are not inflected.
NoProperNouns A detector that filters all proper noun MWEs from the results of its backing detector.
Perfect A detector that finds all of the MWEs in a sentence by using an answer key.
ProperNouns Detects the proper nouns in an ISentence.
ResolveByScore Abstract base class for MWE detector resolvers that use a simple score to resolve conflicts.
SmallestVariance A detector resolver that chooses the MWE with the smallest variance.
StopWords Uses the exhaustive strategy specified in Exhaustive to find the MWEs that consist entirely of stop words in a sentence.
TrulyExhaustive Returns a composite detector backed by the Exhaustive and StopWords detection strategies.
 

Enum Summary
InflectionRule Enumerates the ways in which MWEs with certain parts of speech are allowed to inflect.
 

Package edu.mit.jmwe.detect Description

Provides MWE detector API, a baseline detector, plus numerous other detector implementations.



Copyright © 2011 Massachusetts Institute of Technology. All Rights Reserved.