configs.illinois.Benchmark Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of bat-framework Show documentation

A framework to compare entity annotation systems.

The newest version!

pathToWikiDataSummaryFile=benchmark/systems/illinois/Data/WikiSummary.proto.save

pathToProtobufferWikiAccessIndex=benchmark/systems/illinois/Data/WikiAccessProtoBuffers

pathToSurfaceFormsIndexFile=benchmark/systems/illinois/Data/SurfaceFromsInfo

# the redirects data allows to evaluate results obtained with different versions of Wikipedia
pathToEvaluationRedirectsData=benchmark/systems/illinois/Data/WikiData/RedirectsForEvaluationData/RedirectsFeb2011.txt

pathToPairwiseSemanticInfo=benchmark/systems/illinois/Data/PairwiseSemanticSim/Index

pathToWordNetConfigFile=benchmark/configs/illinois/jwnl_properties_benchmark.xml

pathToTitleCategoryKeywordsInfo=benchmark/systems/illinois/Data/WikiData/TitleSummaryData.txt

pathToModels=benchmark/systems/illinois/Data/Models/DemoModels

pathToStopWords=benchmark/systems/illinois/Data/OtherData/stopwords_big

curatorServerMachine=grandma.cs.uiuc.edu
curatorPort=9010
pathToCuratorCache=benchmark/systems/illinois/Data/TAC_ProblemsAndTextData/CuratorCache

disableCacheUse=true
expectCachedCuratorData=false
bypassTheCurator=true
pathToOfflineNerConfigFile=benchmark/configs/illinois/NER_benchmark.config

disallowEntityOverlap=false
THRES_FOR_ALMOST_UNAMBIGUOUS_SOLUTION=0.99
annotateOnlyAlmostUnambiguousMentions=false
showProgress=true
showMemoryUsage=false 

maxCandidatesToGenerateInitially=20

useOnlyConditionalTitleProbabilitiesInTitleMatchFeatures=false
useGeneratorFeatures=false
useLexicalFeaturesNaive=false
useLexicalFeaturesReweighted=true
useUnambiguousInDisambiguationContext=false
useNamedEntitiesInDisambiguationContext=false
useAllSurfaceFormsInDisambiguationContext=true
useOnlyLinkedSurfaceFormsForDisambiguationContext=true
useCoherenceFeatures=true
useGoldAnnotationsInCoherence=false

generateFeaturesConjunctions=false
numCandidatesAtCoherenceLevel=3

# These are the Edison views and the Curator services used in this configuration
# it's always best to keep these demands on minimum, particularly for the demo,
# since the caching is useless for the demo and the services do tend to break down
NER=true
SHALLOW_PARSE=true
DEPENDENCY=false
POS=false