-
stringRep
java.lang.String stringRep
-
useNGrams
boolean useNGrams
-
conjoinShapeNGrams
boolean conjoinShapeNGrams
-
lowercaseNGrams
boolean lowercaseNGrams
-
dehyphenateNGrams
boolean dehyphenateNGrams
-
usePrev
boolean usePrev
-
useNext
boolean useNext
-
useTags
boolean useTags
-
useWordPairs
boolean useWordPairs
-
useGazettes
boolean useGazettes
-
useSequences
boolean useSequences
-
usePrevSequences
boolean usePrevSequences
-
useNextSequences
boolean useNextSequences
-
useLongSequences
boolean useLongSequences
-
useBoundarySequences
boolean useBoundarySequences
-
useTaggySequences
boolean useTaggySequences
-
useExtraTaggySequences
boolean useExtraTaggySequences
-
dontExtendTaggy
boolean dontExtendTaggy
-
useTaggySequencesShapeInteraction
boolean useTaggySequencesShapeInteraction
-
strictlyZeroethOrder
boolean strictlyZeroethOrder
-
strictlyFirstOrder
boolean strictlyFirstOrder
-
strictlySecondOrder
boolean strictlySecondOrder
-
strictlyThirdOrder
boolean strictlyThirdOrder
-
entitySubclassification
java.lang.String entitySubclassification
-
retainEntitySubclassification
boolean retainEntitySubclassification
-
useGazettePhrases
boolean useGazettePhrases
-
makeConsistent
boolean makeConsistent
-
useViterbi
boolean useViterbi
-
binnedLengths
int[] binnedLengths
-
verboseMode
boolean verboseMode
-
useSum
boolean useSum
-
tolerance
double tolerance
-
printFeatures
java.lang.String printFeatures
-
useSymTags
boolean useSymTags
-
useSymWordPairs
boolean useSymWordPairs
useSymWordPairs Has a small negative effect.
-
printClassifier
java.lang.String printClassifier
-
printClassifierParam
int printClassifierParam
-
intern
boolean intern
-
intern2
boolean intern2
-
selfTest
boolean selfTest
-
sloppyGazette
boolean sloppyGazette
-
cleanGazette
boolean cleanGazette
-
noMidNGrams
boolean noMidNGrams
-
maxNGramLeng
int maxNGramLeng
-
useReverse
boolean useReverse
-
greekifyNGrams
boolean greekifyNGrams
-
useParenMatching
boolean useParenMatching
-
useLemmas
boolean useLemmas
-
usePrevNextLemmas
boolean usePrevNextLemmas
-
normalizeTerms
boolean normalizeTerms
-
normalizeTimex
boolean normalizeTimex
-
useNB
boolean useNB
-
useQN
boolean useQN
-
useFloat
boolean useFloat
-
QNsize
int QNsize
-
QNsize2
int QNsize2
-
maxIterations
int maxIterations
-
wordShape
int wordShape
-
useShapeStrings
boolean useShapeStrings
Set useShapeStrings to be true to say that the model should use word shape features and they are provided in
the tokens, but should not be calculated via a word shape function. This flag must be false if the word shape
features will be calculated; word shape features are also added if there is a defined word shape function.
-
useTypeSeqs
boolean useTypeSeqs
-
useTypeSeqs2
boolean useTypeSeqs2
-
useTypeSeqs3
boolean useTypeSeqs3
-
useDisjunctive
boolean useDisjunctive
-
disjunctionWidth
int disjunctionWidth
-
useDisjunctiveShapeInteraction
boolean useDisjunctiveShapeInteraction
-
useDisjShape
boolean useDisjShape
-
useWord
boolean useWord
-
useClassFeature
boolean useClassFeature
-
useShapeConjunctions
boolean useShapeConjunctions
-
useWordTag
boolean useWordTag
-
useNPHead
boolean useNPHead
-
useNPGovernor
boolean useNPGovernor
-
useHeadGov
boolean useHeadGov
-
useLastRealWord
boolean useLastRealWord
-
useNextRealWord
boolean useNextRealWord
-
useOccurrencePatterns
boolean useOccurrencePatterns
-
useTypeySequences
boolean useTypeySequences
-
justify
boolean justify
-
normalize
boolean normalize
-
priorType
java.lang.String priorType
-
sigma
double sigma
-
epsilon
double epsilon
-
beamSize
int beamSize
-
maxLeft
int maxLeft
-
maxRight
int maxRight
-
usePosition
boolean usePosition
-
useBeginSent
boolean useBeginSent
-
useGazFeatures
boolean useGazFeatures
-
useMoreGazFeatures
boolean useMoreGazFeatures
-
useAbbr
boolean useAbbr
-
useMinimalAbbr
boolean useMinimalAbbr
-
useAbbr1
boolean useAbbr1
-
useMinimalAbbr1
boolean useMinimalAbbr1
-
useMoreAbbr
boolean useMoreAbbr
-
deleteBlankLines
boolean deleteBlankLines
-
useGENIA
boolean useGENIA
-
useTOK
boolean useTOK
-
useABSTR
boolean useABSTR
-
useABSTRFreqDict
boolean useABSTRFreqDict
-
useABSTRFreq
boolean useABSTRFreq
-
useFREQ
boolean useFREQ
-
useABGENE
boolean useABGENE
-
useWEB
boolean useWEB
-
useWEBFreqDict
boolean useWEBFreqDict
-
useIsURL
boolean useIsURL
-
useURLSequences
boolean useURLSequences
-
useIsDateRange
boolean useIsDateRange
-
useEntityTypes
boolean useEntityTypes
-
useEntityTypeSequences
boolean useEntityTypeSequences
-
useEntityRule
boolean useEntityRule
-
useOrdinal
boolean useOrdinal
-
useACR
boolean useACR
-
useANTE
boolean useANTE
-
useMoreTags
boolean useMoreTags
-
useChunks
boolean useChunks
-
useChunkySequences
boolean useChunkySequences
-
usePrevVB
boolean usePrevVB
-
useNextVB
boolean useNextVB
-
useVB
boolean useVB
-
subCWGaz
boolean subCWGaz
-
documentReader
java.lang.String documentReader
-
map
java.lang.String map
-
useWideDisjunctive
boolean useWideDisjunctive
-
wideDisjunctionWidth
int wideDisjunctionWidth
-
useRadical
boolean useRadical
-
useBigramInTwoClique
boolean useBigramInTwoClique
-
morphFeatureFile
java.lang.String morphFeatureFile
-
useReverseAffix
boolean useReverseAffix
-
charHalfWindow
int charHalfWindow
-
useWord1
boolean useWord1
-
useWord2
boolean useWord2
-
useWord3
boolean useWord3
-
useWord4
boolean useWord4
-
useRad1
boolean useRad1
-
useRad2
boolean useRad2
-
useWordn
boolean useWordn
-
useCTBPre1
boolean useCTBPre1
-
useCTBSuf1
boolean useCTBSuf1
-
useASBCPre1
boolean useASBCPre1
-
useASBCSuf1
boolean useASBCSuf1
-
usePKPre1
boolean usePKPre1
-
usePKSuf1
boolean usePKSuf1
-
useHKPre1
boolean useHKPre1
-
useHKSuf1
boolean useHKSuf1
-
useCTBChar2
boolean useCTBChar2
-
useASBCChar2
boolean useASBCChar2
-
useHKChar2
boolean useHKChar2
-
usePKChar2
boolean usePKChar2
-
useRule2
boolean useRule2
-
useDict2
boolean useDict2
-
useOutDict2
boolean useOutDict2
-
outDict2
java.lang.String outDict2
-
useDictleng
boolean useDictleng
-
useDictCTB2
boolean useDictCTB2
-
useDictASBC2
boolean useDictASBC2
-
useDictPK2
boolean useDictPK2
-
useDictHK2
boolean useDictHK2
-
useBig5
boolean useBig5
-
useNegDict2
boolean useNegDict2
-
useNegDict3
boolean useNegDict3
-
useNegDict4
boolean useNegDict4
-
useNegCTBDict2
boolean useNegCTBDict2
-
useNegCTBDict3
boolean useNegCTBDict3
-
useNegCTBDict4
boolean useNegCTBDict4
-
useNegASBCDict2
boolean useNegASBCDict2
-
useNegASBCDict3
boolean useNegASBCDict3
-
useNegASBCDict4
boolean useNegASBCDict4
-
useNegHKDict2
boolean useNegHKDict2
-
useNegHKDict3
boolean useNegHKDict3
-
useNegHKDict4
boolean useNegHKDict4
-
useNegPKDict2
boolean useNegPKDict2
-
useNegPKDict3
boolean useNegPKDict3
-
useNegPKDict4
boolean useNegPKDict4
-
usePre
boolean usePre
-
useSuf
boolean useSuf
-
useRule
boolean useRule
-
useHk
boolean useHk
-
useMsr
boolean useMsr
-
useMSRChar2
boolean useMSRChar2
-
usePk
boolean usePk
-
useAs
boolean useAs
-
useFilter
boolean useFilter
-
largeChSegFile
boolean largeChSegFile
-
useRad2b
boolean useRad2b
-
keepEnglishWhitespaces
boolean keepEnglishWhitespaces
Keep the whitespace between English words in testFile when printing out
answers. Doesn't really change the content of the CoreLabels. (For Chinese
segmentation.)
-
keepAllWhitespaces
boolean keepAllWhitespaces
Keep all the whitespace words in testFile when printing out answers.
Doesn't really change the content of the CoreLabels. (For Chinese
segmentation.)
-
sighanPostProcessing
boolean sighanPostProcessing
-
useChPos
boolean useChPos
use POS information (an "open" feature for Chinese segmentation)
-
normalizationTable
java.lang.String normalizationTable
-
dictionary
java.lang.String dictionary
-
serializedDictionary
java.lang.String serializedDictionary
-
dictionary2
java.lang.String dictionary2
-
normTableEncoding
java.lang.String normTableEncoding
-
sighanCorporaDict
java.lang.String sighanCorporaDict
for Sighan bakeoff 2005, the path to the dictionary of bigrams appeared in
corpus
-
useWordShapeGaz
boolean useWordShapeGaz
-
wordShapeGaz
java.lang.String wordShapeGaz
-
splitDocuments
boolean splitDocuments
-
printXML
boolean printXML
-
useSeenFeaturesOnly
boolean useSeenFeaturesOnly
-
lastNameList
java.lang.String lastNameList
-
maleNameList
java.lang.String maleNameList
-
femaleNameList
java.lang.String femaleNameList
-
inputEncoding
java.lang.String inputEncoding
-
bioSubmitOutput
boolean bioSubmitOutput
-
numRuns
int numRuns
-
answerFile
java.lang.String answerFile
-
altAnswerFile
java.lang.String altAnswerFile
-
dropGaz
java.lang.String dropGaz
-
printGazFeatures
java.lang.String printGazFeatures
-
numStartLayers
int numStartLayers
-
dump
boolean dump
-
mergeTags
boolean mergeTags
-
splitOnHead
boolean splitOnHead
-
featureCountThreshold
int featureCountThreshold
-
featureWeightThreshold
double featureWeightThreshold
-
featureFactory
java.lang.String featureFactory
-
featureFactoryArgs
java.lang.Object[] featureFactoryArgs
-
backgroundSymbol
java.lang.String backgroundSymbol
-
useObservedSequencesOnly
boolean useObservedSequencesOnly
-
maxDocSize
int maxDocSize
-
printProbs
boolean printProbs
-
printFirstOrderProbs
boolean printFirstOrderProbs
-
saveFeatureIndexToDisk
boolean saveFeatureIndexToDisk
-
removeBackgroundSingletonFeatures
boolean removeBackgroundSingletonFeatures
-
doGibbs
boolean doGibbs
-
numSamples
int numSamples
-
useNERPrior
boolean useNERPrior
-
useAcqPrior
boolean useAcqPrior
-
useUniformPrior
boolean useUniformPrior
-
useMUCFeatures
boolean useMUCFeatures
-
annealingRate
double annealingRate
-
annealingType
java.lang.String annealingType
-
loadProcessedData
java.lang.String loadProcessedData
-
initViterbi
boolean initViterbi
-
useUnknown
boolean useUnknown
-
checkNameList
boolean checkNameList
-
useSemPrior
boolean useSemPrior
-
useFirstWord
boolean useFirstWord
-
useNumberFeature
boolean useNumberFeature
-
ocrFold
int ocrFold
-
classifierType
java.lang.String classifierType
-
svmModelFile
java.lang.String svmModelFile
-
inferenceType
java.lang.String inferenceType
-
useLemmaAsWord
boolean useLemmaAsWord
-
type
java.lang.String type
-
readerAndWriter
java.lang.String readerAndWriter
-
comboProps
java.util.List<E> comboProps
-
usePrediction
boolean usePrediction
-
useAltGazFeatures
boolean useAltGazFeatures
-
gazFilesFile
java.lang.String gazFilesFile
-
usePrediction2
boolean usePrediction2
-
baseTrainDir
java.lang.String baseTrainDir
-
baseTestDir
java.lang.String baseTestDir
-
trainFiles
java.lang.String trainFiles
A regex pattern for files, which will be evaluated within a particular directory.
If non-null, used over trainFileList and trainFile.
-
trainFileList
java.lang.String trainFileList
-
testFiles
java.lang.String testFiles
-
trainDirs
java.lang.String trainDirs
-
testDirs
java.lang.String testDirs
-
useOnlySeenWeights
boolean useOnlySeenWeights
-
predProp
java.lang.String predProp
-
pad
CoreLabel pad
-
useObservedFeaturesOnly
boolean useObservedFeaturesOnly
-
distSimLexicon
java.lang.String distSimLexicon
-
useDistSim
boolean useDistSim
-
removeTopN
int removeTopN
-
numTimesRemoveTopN
int numTimesRemoveTopN
-
randomizedRatio
double randomizedRatio
-
removeTopNPercent
double removeTopNPercent
-
purgeFeatures
int purgeFeatures
-
booleanFeatures
boolean booleanFeatures
-
iobWrapper
boolean iobWrapper
-
iobTags
boolean iobTags
-
useSegmentation
boolean useSegmentation
Binary segmentation feature for character-based Chinese NER.
-
memoryThrift
boolean memoryThrift
-
timitDatum
boolean timitDatum
-
serializeDatasetsDir
java.lang.String serializeDatasetsDir
-
loadDatasetsDir
java.lang.String loadDatasetsDir
-
pushDir
java.lang.String pushDir
-
purgeDatasets
boolean purgeDatasets
-
keepOBInMemory
boolean keepOBInMemory
-
fakeDataset
boolean fakeDataset
-
restrictTransitionsTimit
boolean restrictTransitionsTimit
-
numDatasetsPerFile
int numDatasetsPerFile
-
useTitle
boolean useTitle
-
lowerNewgeneThreshold
boolean lowerNewgeneThreshold
-
useEitherSideWord
boolean useEitherSideWord
-
useEitherSideDisjunctive
boolean useEitherSideDisjunctive
-
twoStage
boolean twoStage
-
crfType
java.lang.String crfType
-
featureThreshold
int featureThreshold
-
featThreshFile
java.lang.String featThreshFile
-
featureDiffThresh
double featureDiffThresh
-
numTimesPruneFeatures
int numTimesPruneFeatures
-
newgeneThreshold
double newgeneThreshold
-
doAdaptation
boolean doAdaptation
-
useInternal
boolean useInternal
-
useExternal
boolean useExternal
-
selfTrainConfidenceThreshold
double selfTrainConfidenceThreshold
-
selfTrainIterations
int selfTrainIterations
-
selfTrainWindowSize
int selfTrainWindowSize
-
useHuber
boolean useHuber
-
useQuartic
boolean useQuartic
-
adaptSigma
double adaptSigma
-
numFolds
int numFolds
-
startFold
int startFold
-
endFold
int endFold
-
cacheNGrams
boolean cacheNGrams
-
outputFormat
java.lang.String outputFormat
-
useSMD
boolean useSMD
-
useSGDtoQN
boolean useSGDtoQN
-
useStochasticQN
boolean useStochasticQN
-
useScaledSGD
boolean useScaledSGD
-
scaledSGDMethod
int scaledSGDMethod
-
SGDPasses
int SGDPasses
-
QNPasses
int QNPasses
-
tuneSGD
boolean tuneSGD
-
stochasticMethod
StochasticCalculateMethods stochasticMethod
-
initialGain
double initialGain
-
stochasticBatchSize
int stochasticBatchSize
-
useSGD
boolean useSGD
-
gainSGD
double gainSGD
-
useHybrid
boolean useHybrid
-
hybridCutoffIteration
int hybridCutoffIteration
-
outputIterationsToFile
boolean outputIterationsToFile
-
testObjFunction
boolean testObjFunction
-
testVariance
boolean testVariance
-
SGD2QNhessSamples
int SGD2QNhessSamples
-
testHessSamples
boolean testHessSamples
-
CRForder
int CRForder
-
CRFwindow
int CRFwindow
-
estimateInitial
boolean estimateInitial
-
outputEncoding
java.lang.String outputEncoding
-
useKBest
boolean useKBest
-
searchGraphPrefix
java.lang.String searchGraphPrefix
-
searchGraphPrune
double searchGraphPrune
-
kBest
int kBest
-
useFeaturesC4gram
boolean useFeaturesC4gram
-
useFeaturesC5gram
boolean useFeaturesC5gram
-
useFeaturesC6gram
boolean useFeaturesC6gram
-
useFeaturesCpC4gram
boolean useFeaturesCpC4gram
-
useFeaturesCpC5gram
boolean useFeaturesCpC5gram
-
useFeaturesCpC6gram
boolean useFeaturesCpC6gram
-
useUnicodeType
boolean useUnicodeType
-
useUnicodeType4gram
boolean useUnicodeType4gram
-
useUnicodeType5gram
boolean useUnicodeType5gram
-
use4Clique
boolean use4Clique
-
useUnicodeBlock
boolean useUnicodeBlock
-
useShapeStrings1
boolean useShapeStrings1
-
useShapeStrings3
boolean useShapeStrings3
-
useShapeStrings4
boolean useShapeStrings4
-
useShapeStrings5
boolean useShapeStrings5
-
useGoodForNamesCpC
boolean useGoodForNamesCpC
-
useDictionaryConjunctions
boolean useDictionaryConjunctions
-
expandMidDot
boolean expandMidDot
-
printFeaturesUpto
int printFeaturesUpto
-
useDictionaryConjunctions3
boolean useDictionaryConjunctions3
-
useWordUTypeConjunctions2
boolean useWordUTypeConjunctions2
-
useWordUTypeConjunctions3
boolean useWordUTypeConjunctions3
-
useWordShapeConjunctions2
boolean useWordShapeConjunctions2
-
useWordShapeConjunctions3
boolean useWordShapeConjunctions3
-
useMidDotShape
boolean useMidDotShape
-
augmentedDateChars
boolean augmentedDateChars
-
suppressMidDotPostprocessing
boolean suppressMidDotPostprocessing
-
printNR
boolean printNR
-
classBias
java.lang.String classBias
-
printLabelValue
boolean printLabelValue
-
useRobustQN
boolean useRobustQN
-
combo
boolean combo
-
useGenericFeatures
boolean useGenericFeatures
-
verboseForTrueCasing
boolean verboseForTrueCasing
-
trainHierarchical
java.lang.String trainHierarchical
-
domain
java.lang.String domain
-
baseline
boolean baseline
-
transferSigmas
java.lang.String transferSigmas
-
doFE
boolean doFE
-
restrictLabels
boolean restrictLabels
-
announceObjectBankEntries
boolean announceObjectBankEntries
-
l1reg
double l1reg
-
mixedCaseMapFile
java.lang.String mixedCaseMapFile
-
auxTrueCaseModels
java.lang.String auxTrueCaseModels
-
use2W
boolean use2W
-
useLC
boolean useLC
-
useYetMoreCpCShapes
boolean useYetMoreCpCShapes
-
useIfInteger
boolean useIfInteger
-
exportFeatures
java.lang.String exportFeatures
-
useInPlaceSGD
boolean useInPlaceSGD
-
useTopics
boolean useTopics
-
evaluateIters
int evaluateIters
-
evalCmd
java.lang.String evalCmd
-
evaluateTrain
boolean evaluateTrain
-
tuneSampleSize
int tuneSampleSize
-
usePhraseFeatures
boolean usePhraseFeatures
-
usePhraseWords
boolean usePhraseWords
-
usePhraseWordTags
boolean usePhraseWordTags
-
usePhraseWordSpecialTags
boolean usePhraseWordSpecialTags
-
useCommonWordsFeature
boolean useCommonWordsFeature
-
useProtoFeatures
boolean useProtoFeatures
-
useWordnetFeatures
boolean useWordnetFeatures
-
tokenFactory
java.lang.String tokenFactory
-
tokenFactoryArgs
java.lang.Object[] tokenFactoryArgs
-
tokensAnnotationClassName
java.lang.String tokensAnnotationClassName
-
useCorefFeatures
boolean useCorefFeatures
-
wikiFeatureDbFile
java.lang.String wikiFeatureDbFile
-
useNoisyNonNoisyFeature
boolean useNoisyNonNoisyFeature
-
useYear
boolean useYear
-
useSentenceNumber
boolean useSentenceNumber
-
useLabelSource
boolean useLabelSource
-
casedDistSim
boolean casedDistSim
Whether to (not) lowercase tokens before looking them up in distsim
lexicon. By default lowercasing was done, but now it doesn't have to be
true :-).
-
distSimFileFormat
java.lang.String distSimFileFormat
The format of the distsim file. Known values are: alexClark = TSV file.
word TAB clusterNumber [optional other content] terryKoo = TSV file.
clusterBitString TAB word TAB frequency
-
distSimMaxBits
int distSimMaxBits
If this number is greater than 0, the distSim class is assume to be a bit
string and is truncated at this many characters. Normal distSim features
will then use this amount of resolution. Extra, special distsim features
may work at a coarser level of resolution. Since the lexicon only stores
this length of bit string, there is then no way to have finer-grained
clusters.
-
numberEquivalenceDistSim
boolean numberEquivalenceDistSim
If this is set to true, all digit characters get mapped to '9' in a distsim
lexicon and for lookup. This is a simple word shaping that can shrink
distsim lexicons and improve their performance.
-
unknownWordDistSimClass
java.lang.String unknownWordDistSimClass
What class to assign to words not found in the dist sim lexicon. You might
want to make it a known class, if one is the "default class.
-
useNeighborNGrams
boolean useNeighborNGrams
Use prefixes and suffixes from the previous and current word in edge clique.
-
wordFunction
java.util.function.Function<T,R> wordFunction
This function maps words in the training or test data to new
words. They are used at the feature extractor level, ie in the
FeatureFactory. For now, only the NERFeatureFactory uses this.
-
plainTextDocumentReaderAndWriter
java.lang.String plainTextDocumentReaderAndWriter
-
useBagOfWords
boolean useBagOfWords
Use a bag of all words as a feature. Perhaps this will find some
words that indicate certain types of entities are present.
-
evaluateBackground
boolean evaluateBackground
When scoring, count the background symbol stats too. Useful for
things where the background symbol is particularly meaningful,
such as truecase.
-
numLopExpert
int numLopExpert
Number of experts to be used in Logarithmic Opinion Pool (product of experts) training
default value is 1
-
includeFullCRFInLOP
boolean includeFullCRFInLOP
-
backpropLopTraining
boolean backpropLopTraining
-
randomLopWeights
boolean randomLopWeights
-
randomLopFeatureSplit
boolean randomLopFeatureSplit
-
nonLinearCRF
boolean nonLinearCRF
-
secondOrderNonLinear
boolean secondOrderNonLinear
-
numHiddenUnits
int numHiddenUnits
-
useOutputLayer
boolean useOutputLayer
-
useHiddenLayer
boolean useHiddenLayer
-
gradientDebug
boolean gradientDebug
-
checkGradient
boolean checkGradient
-
useSigmoid
boolean useSigmoid
-
skipOutputRegularization
boolean skipOutputRegularization
-
sparseOutputLayer
boolean sparseOutputLayer
-
tieOutputLayer
boolean tieOutputLayer
-
blockInitialize
boolean blockInitialize
-
softmaxOutputLayer
boolean softmaxOutputLayer
-
loadBisequenceClassifierEn
java.lang.String loadBisequenceClassifierEn
Bisequence CRF parameters
-
loadBisequenceClassifierCh
java.lang.String loadBisequenceClassifierCh
-
bisequenceClassifierPropEn
java.lang.String bisequenceClassifierPropEn
-
bisequenceClassifierPropCh
java.lang.String bisequenceClassifierPropCh
-
bisequenceTestFileEn
java.lang.String bisequenceTestFileEn
-
bisequenceTestFileCh
java.lang.String bisequenceTestFileCh
-
bisequenceTestOutputEn
java.lang.String bisequenceTestOutputEn
-
bisequenceTestOutputCh
java.lang.String bisequenceTestOutputCh
-
bisequenceTestAlignmentFile
java.lang.String bisequenceTestAlignmentFile
-
bisequenceAlignmentTestOutput
java.lang.String bisequenceAlignmentTestOutput
-
bisequencePriorType
int bisequencePriorType
-
bisequenceAlignmentPriorPenaltyCh
java.lang.String bisequenceAlignmentPriorPenaltyCh
-
bisequenceAlignmentPriorPenaltyEn
java.lang.String bisequenceAlignmentPriorPenaltyEn
-
alignmentPruneThreshold
double alignmentPruneThreshold
-
alignmentDecodeThreshold
double alignmentDecodeThreshold
-
factorInAlignmentProb
boolean factorInAlignmentProb
-
useChromaticSampling
boolean useChromaticSampling
-
useSequentialScanSampling
boolean useSequentialScanSampling
-
maxAllowedChromaticSize
int maxAllowedChromaticSize
-
keepEmptySentences
boolean keepEmptySentences
Whether or not to keep blank sentences when processing. Useful
for systems such as the segmenter if you want to line up each
line exactly, including blank lines.
-
useBilingualNERPrior
boolean useBilingualNERPrior
-
samplingSpeedUpThreshold
int samplingSpeedUpThreshold
-
entityMatrixCh
java.lang.String entityMatrixCh
-
entityMatrixEn
java.lang.String entityMatrixEn
-
multiThreadGibbs
int multiThreadGibbs
-
matchNERIncentive
boolean matchNERIncentive
-
useEmbedding
boolean useEmbedding
-
prependEmbedding
boolean prependEmbedding
-
embeddingWords
java.lang.String embeddingWords
-
embeddingVectors
java.lang.String embeddingVectors
-
transitionEdgeOnly
boolean transitionEdgeOnly
-
priorLambda
double priorLambda
-
addCapitalFeatures
boolean addCapitalFeatures
-
arbitraryInputLayerSize
int arbitraryInputLayerSize
-
noEdgeFeature
boolean noEdgeFeature
-
terminateOnEvalImprovement
boolean terminateOnEvalImprovement
-
terminateOnEvalImprovementNumOfEpoch
int terminateOnEvalImprovementNumOfEpoch
-
useMemoryEvaluator
boolean useMemoryEvaluator
-
suppressTestDebug
boolean suppressTestDebug
-
useOWLQN
boolean useOWLQN
-
printWeights
boolean printWeights
-
totalDataSlice
int totalDataSlice
-
numOfSlices
int numOfSlices
-
regularizeSoftmaxTieParam
boolean regularizeSoftmaxTieParam
-
softmaxTieLambda
double softmaxTieLambda
-
totalFeatureSlice
int totalFeatureSlice
-
numOfFeatureSlices
int numOfFeatureSlices
-
addBiasToEmbedding
boolean addBiasToEmbedding
-
hardcodeSoftmaxOutputWeights
boolean hardcodeSoftmaxOutputWeights
-
useNERPriorBIO
boolean useNERPriorBIO
-
entityMatrix
java.lang.String entityMatrix
-
multiThreadClassifier
int multiThreadClassifier
-
useDualDecomp
boolean useDualDecomp
-
biAlignmentPriorIsPMI
boolean biAlignmentPriorIsPMI
-
dampDDStepSizeWithAlignmentProb
boolean dampDDStepSizeWithAlignmentProb
-
dualDecompAlignment
boolean dualDecompAlignment
-
dualDecompInitialStepSizeAlignment
double dualDecompInitialStepSizeAlignment
-
dualDecompNotBIO
boolean dualDecompNotBIO
-
berkeleyAlignerLoadPath
java.lang.String berkeleyAlignerLoadPath
-
useBerkeleyAlignerForViterbi
boolean useBerkeleyAlignerForViterbi
-
useBerkeleyCompetitivePosterior
boolean useBerkeleyCompetitivePosterior
-
useDenero
boolean useDenero
-
alignDDAlpha
double alignDDAlpha
-
factorInBiEdgePotential
boolean factorInBiEdgePotential
-
noNeighborConstraints
boolean noNeighborConstraints
-
includeC2EViterbi
boolean includeC2EViterbi
-
initWithPosterior
boolean initWithPosterior
-
nerSkipFirstK
int nerSkipFirstK
-
nerSlowerTimes
int nerSlowerTimes
-
powerAlignProb
boolean powerAlignProb
-
powerAlignProbAsAddition
boolean powerAlignProbAsAddition
-
initWithNERPosterior
boolean initWithNERPosterior
-
applyNERPenalty
boolean applyNERPenalty
-
printFactorTable
boolean printFactorTable
-
useAdaGradFOBOS
boolean useAdaGradFOBOS
-
initRate
double initRate
-
groupByFeatureTemplate
boolean groupByFeatureTemplate
-
groupByOutputClass
boolean groupByOutputClass
-
priorAlpha
double priorAlpha
-
splitWordRegex
java.lang.String splitWordRegex
-
groupByInput
boolean groupByInput
-
groupByHiddenUnit
boolean groupByHiddenUnit
-
unigramLM
java.lang.String unigramLM
-
bigramLM
java.lang.String bigramLM
-
wordSegBeamSize
int wordSegBeamSize
-
vocabFile
java.lang.String vocabFile
-
normalizedFile
java.lang.String normalizedFile
-
averagePerceptron
boolean averagePerceptron
-
loadCRFSegmenterPath
java.lang.String loadCRFSegmenterPath
-
loadPCTSegmenterPath
java.lang.String loadPCTSegmenterPath
-
crfSegmenterProp
java.lang.String crfSegmenterProp
-
pctSegmenterProp
java.lang.String pctSegmenterProp
-
intermediateSegmenterOut
java.lang.String intermediateSegmenterOut
-
intermediateSegmenterModel
java.lang.String intermediateSegmenterModel
-
dualDecompMaxItr
int dualDecompMaxItr
-
dualDecompInitialStepSize
double dualDecompInitialStepSize
-
dualDecompDebug
boolean dualDecompDebug
-
useCWSWordFeatures
boolean useCWSWordFeatures
-
useCWSWordFeaturesAll
boolean useCWSWordFeaturesAll
-
useCWSWordFeaturesBigram
boolean useCWSWordFeaturesBigram
-
pctSegmenterLenAdjust
boolean pctSegmenterLenAdjust
-
useTrainLexicon
boolean useTrainLexicon
-
useCWSFeatures
boolean useCWSFeatures
-
appendLC
boolean appendLC
-
perceptronDebug
boolean perceptronDebug
-
pctSegmenterScaleByCRF
boolean pctSegmenterScaleByCRF
-
pctSegmenterScale
double pctSegmenterScale
-
separateASCIIandRange
boolean separateASCIIandRange
-
dropoutRate
double dropoutRate
-
dropoutScale
double dropoutScale
-
multiThreadGrad
int multiThreadGrad
-
maxQNItr
int maxQNItr
-
dropoutApprox
boolean dropoutApprox
-
unsupDropoutFile
java.lang.String unsupDropoutFile
-
unsupDropoutScale
double unsupDropoutScale
-
startEvaluateIters
int startEvaluateIters
-
multiThreadPerceptron
int multiThreadPerceptron
-
lazyUpdate
boolean lazyUpdate
-
featureCountThresh
int featureCountThresh
-
geDebug
boolean geDebug
-
doFeatureDiscovery
boolean doFeatureDiscovery
-
learnCHBasedOnEN
boolean learnCHBasedOnEN
-
learnENBasedOnCH
boolean learnENBasedOnCH
-
loadWeightsFromEN
java.lang.String loadWeightsFromEN
-
loadWeightsFromCH
java.lang.String loadWeightsFromCH
-
serializeToEN
java.lang.String serializeToEN
-
serializeToCH
java.lang.String serializeToCH
-
testFileEN
java.lang.String testFileEN
-
testFileCH
java.lang.String testFileCH
-
unsupFileEN
java.lang.String unsupFileEN
-
unsupFileCH
java.lang.String unsupFileCH
-
unsupAlignFile
java.lang.String unsupAlignFile
-
supFileEN
java.lang.String supFileEN
-
supFileCH
java.lang.String supFileCH
-
loadFeatureIndexFromEN
java.lang.String loadFeatureIndexFromEN
-
loadFeatureIndexFromCH
java.lang.String loadFeatureIndexFromCH
-
lambdaEN
double lambdaEN
-
lambdaCH
double lambdaCH
-
alternateTraining
boolean alternateTraining
-
weightByEntropy
boolean weightByEntropy
-
useKL
boolean useKL
-
useHardGE
boolean useHardGE
-
useCRFforUnsup
boolean useCRFforUnsup
-
useGEforSup
boolean useGEforSup
-
useKnownLCWords
boolean useKnownLCWords
-
featureFactories
java.lang.String[] featureFactories
-
featureFactoriesArgs
java.util.List<E> featureFactoriesArgs
-
useNoisyLabel
boolean useNoisyLabel
-
errorMatrix
java.lang.String errorMatrix
-
printTrainLabels
boolean printTrainLabels
-
labelDictionaryCutoff
int labelDictionaryCutoff
-
useAdaDelta
boolean useAdaDelta
-
useAdaDiff
boolean useAdaDiff
-
adaGradEps
double adaGradEps
-
adaDeltaRho
double adaDeltaRho
-
useRandomSeed
boolean useRandomSeed
-
terminateOnAvgImprovement
boolean terminateOnAvgImprovement
-
strictGoodCoNLL
boolean strictGoodCoNLL
-
removeStrictGoodCoNLLDuplicates
boolean removeStrictGoodCoNLLDuplicates
-
priorModelFactory
java.lang.String priorModelFactory
A class name for a factory that vends a prior NER model that
implements both SequenceModel and SequenceListener, and which
is used in the Gibbs sampling sequence model inference.
-
useUndirectedDisjunctive
boolean useUndirectedDisjunctive
Put in undirected (left/right) bag of words features for local
neighborhood. Seems much worse than regular useDisjunctive.
-
splitSlashHyphenWords
boolean splitSlashHyphenWords
-
maxAdditionalKnownLCWords
int maxAdditionalKnownLCWords
How many words it is okay to add to knownLCWords after initial training.
If this number is negative, then add any number of further words during classifying/testing.
If this number is non-negative (greater than or equal to 0), then add at most this many words
to the knownLCWords. By default, this is now set to 0, so there is no transductive learning on the
test set, since too many people complained about results changing over runs. However, traditionally
we used a non-zero value, and this usually helps performance a bit (until 2014 it was -1, then it
was set to 10_000, so that memory would not grow without bound if a SequenceClassifier is run for
a long time.
-
slashHyphenTreatment
SeqClassifierFlags.SlashHyphenEnum slashHyphenTreatment
-
useTitle2
boolean useTitle2
-
showNCCInfo
boolean showNCCInfo
-
showCCInfo
boolean showCCInfo
-
crfToExamine
java.lang.String crfToExamine
-
useSUTime
boolean useSUTime
-
applyNumericClassifiers
boolean applyNumericClassifiers
-
combinationMode
java.lang.String combinationMode
-
nerModel
java.lang.String nerModel
-
useMoreNeighborNGrams
boolean useMoreNeighborNGrams
Use prefixes and suffixes from the previous and next word in node clique.