• Home
• ai.tock
• tock-nlp-model-stanford
• 23.3.4
• source code
• crfclassifier.properties

×

Project download

Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $

You can buy this project and download/modify it how often you want.

stanford.crfclassifier.properties Maven / Gradle / Ivy

tolerance=1e-4
map=word=0,answer=1
useGenericFeatures=true
useWord=true
useNGrams=true
#Better results without mid words features -> noMidNGrams=true
noMidNGrams=true
#Optimum for maxNGramLeng is between 15 and 18
maxNGramLeng=18
usePrev=true
useNext=true
useSequences=true
usePrevSequences=true
maxLeft=1
useTypeSeqs=true
useTypeSeqs2=true
useTypeySequences=true

# better performance and same accuracy when true
useObservedSequencesOnly=true

#same performance and same/worse results when true
saveFeatureIndexToDisk=false

# almost same performance and better results when true
lowercaseNGrams=true

#same accuracy, slightly better performance when true
dehyphenateNGrams=true

#worse performance and same/worse accuracy when true
conjoinShapeNGrams=false

#worse accuracy and way worse performance when true
useNeighborNGrams=false

# slightly worse performance but better accuracy when true
useTags=true

# better results and slightly better performance when true
useNextSequences=true

# same results with worse performance when true
useLongSequences=false

# performance = / accuracy - when true
useSymTags=false

# performance = / accuracy = when true
cacheNGrams=false

# performance = / accuracy = when true
useLemmas=false

# performance - / accuracy = when true
usePrevNextLemmas=false

# performance = / accuracy = when true
useLemmaAsWord=false

#same accuracy worse performance when true
useWordPairs=false
wordShape=chris2useLC
QNsize=10
QNsize2=10