GitHub - vijaydone/prDetect: Paraphrase detection using supervised machine learning

vijaydone / prDetect Public

forked from jaisong87/prDetect

Notifications You must be signed in to change notification settings
Fork 0
Star 0

Paraphrase detection using supervised machine learning

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
Analytics		Analytics
Content		Content
Curl		Curl
FeatureExtract		FeatureExtract
LATEST_NgrFiles		LATEST_NgrFiles
LatestFeatures		LatestFeatures
Preprocess		Preprocess
oldFiles		oldFiles
.lowerCaser.cpp.swp		.lowerCaser.cpp.swp
.wekaFeed.cpp.swp		.wekaFeed.cpp.swp
1		1
AJFindingScoreForAll.cpp		AJFindingScoreForAll.cpp
LCSFeature		LCSFeature
LunigramFeature		LunigramFeature
LunigramFeature.cpp		LunigramFeature.cpp
LunigramFetch		LunigramFetch
NgramGen		NgramGen
OLDUNIGGEN		OLDUNIGGEN
PLAINSIMILARITYSCORE		PLAINSIMILARITYSCORE
PNNFeature		PNNFeature
POSBigramFetch		POSBigramFetch
POSBigramFetch.cpp		POSBigramFetch.cpp
POSBigramFetcher		POSBigramFetcher
POSSkipGramFeature		POSSkipGramFeature
POSSkipGramFetch		POSSkipGramFetch
POSSkipGramFetch.cpp		POSSkipGramFetch.cpp
POSSkipGramFetcher		POSSkipGramFetcher
POSTrigramFetch		POSTrigramFetch
POSTrigramFetch.cpp		POSTrigramFetch.cpp
POSbigramFeature		POSbigramFeature
POStrigramFeature		POStrigramFeature
ParaResults		ParaResults
ParaSen1		ParaSen1
ParaSen2		ParaSen2
PosNeg		PosNeg
PosNegNeutal		PosNegNeutal
README		README
STEMMEDUNIGRAMGEN		STEMMEDUNIGRAMGEN
SVOagreement		SVOagreement
SVOscore1Feature		SVOscore1Feature
SVOscore2Feature		SVOscore2Feature
Samples		Samples
SbigramFetch.cpp		SbigramFetch.cpp
StemmedParaSen1		StemmedParaSen1
StemmedParaSen2		StemmedParaSen2
StemmedUnigram.cpp		StemmedUnigram.cpp
StrippedParaSen1		StrippedParaSen1
StrippedParaSen2		StrippedParaSen2
WEKAFEED		WEKAFEED
a.out		a.out
bigramFeature		bigramFeature
bigramFetch		bigramFetch
bigramFetch.cpp		bigramFetch.cpp
bigramSentenceGen.cpp		bigramSentenceGen.cpp
feature.arff		feature.arff
ftest.dat		ftest.dat
ftest1.dat		ftest1.dat
ftest2.dat		ftest2.dat
ftrain.dat		ftrain.dat
getPolarityDis		getPolarityDis
getPolarityDis2		getPolarityDis2
intelligentSVOED		intelligentSVOED
intelligentsynonymFeature		intelligentsynonymFeature
intellingentSVO1		intellingentSVO1
intellingentSVO2		intellingentSVO2
lcsFetch		lcsFetch
lcsFetch.cpp		lcsFetch.cpp
misclasses.txt		misclasses.txt
nGram.cpp		nGram.cpp
nGramFeature		nGramFeature
nGramFeature~		nGramFeature~
oldUnigramFeature		oldUnigramFeature
oldUnigramFeature.cpp		oldUnigramFeature.cpp
output		output
plainSimilarity.cpp		plainSimilarity.cpp
plainSimilarityFeature		plainSimilarityFeature
polarityDis.cpp		polarityDis.cpp
polarityDis2.cpp		polarityDis2.cpp
polarityDisFeature		polarityDisFeature
polarityDisFeature2		polarityDisFeature2
properNounMatcher		properNounMatcher
properNounMatcher.cpp		properNounMatcher.cpp
sBigramFetch		sBigramFetch
skipGram.cpp		skipGram.cpp
skipGram.h		skipGram.h
skipGram.h.gch		skipGram.h.gch
skipGramFeature		skipGramFeature
skipGramFetch		skipGramFetch
stemSimilarityFeature		stemSimilarityFeature
synonymFeature		synonymFeature
tmp.cpp		tmp.cpp
trainResults		trainResults
trainS1		trainS1
trainS1POS		trainS1POS
trainS1Tagged		trainS1Tagged
trainS2		trainS2
trainS2POS		trainS2POS
trainS2Tagged		trainS2Tagged
trigramFeature		trigramFeature
trigramFetch		trigramFetch

Repository files navigation

This project is for Paraphrase detection using Microsoft Research Corpus.
The code gives around 83.1% which is better than state of the art. A lot
of cleanup is required and some new features have to be incorporated for
better results. This would be added soon.