cmu.arktweetnlp.impl.features

Interface FeatureExtractor.FeatureExtractorInterface

All Known Implementing Classes:

MetaphoneFeatures.MetaphoneLexical, MiscFeatures.CapitalizationFeatures, MiscFeatures.Next2Words, MiscFeatures.NextWord, MiscFeatures.NgramPrefix, MiscFeatures.NgramSuffix, MiscFeatures.Positions, MiscFeatures.Prev2Words, MiscFeatures.PrevNext, MiscFeatures.PrevWord, MiscFeatures.SimpleOrthFeatures, MiscFeatures.URLFeatures, MiscFeatures.WordformFeatures, WordClusterPaths, WordListFeatures.Listofnames, WordListFeatures.MetaphonePOSDict, WordListFeatures.POSTagDict

Enclosing class:

FeatureExtractor

public static interface FeatureExtractor.FeatureExtractorInterface

Method Summary

Methods

Modifier and Type	Method and Description
void	addFeatures(java.util.List<java.lang.String> tokens, FeatureExtractor.PositionFeaturePairs positionFeaturePairs) Input: sentence Output: labelIndexes, featureIDs/Values through positionFeaturePairs We want to yield a sequence of (t, featID, featValue) pairs, to be conjuncted against label IDs at position t.

Method Detail

addFeatures

void addFeatures(java.util.List<java.lang.String> tokens,
               FeatureExtractor.PositionFeaturePairs positionFeaturePairs)

Input: sentence Output: labelIndexes, featureIDs/Values through positionFeaturePairs We want to yield a sequence of (t, featID, featValue) pairs, to be conjuncted against label IDs at position t. Represent as parallel arrays. Ick yes, but we want to save object allocations (is this crazy?) This method should append to them.