GATE功能介绍(对外).pptVIP

概览（1）GATEisaGeneralArchitectureforTextEngineeringDevelopedbytheNaturalLanguageProcessingResearchGroupwithintheDepartmentofComputerScienceattheUniversityofSheffield概览（2）LanguageResources(LRs) referstodata-onlyresourcessuchasdocument,corpus.ProcessingResources(PRs) referstoresourceswhosecharacterisprincipallyprogrammaticoralgorithmic,suchastokeniser,POStagger.Applications modelacontrolstrategyfortheexecutionofPRs. Therearetwomaintypesofpipeline:SimplepipelinesCorpuspipelines概览（4）功能介绍Tokeniser 实现分词功能，每个Token标注包括的属性有： kind:Word,Number,Symbol,Punctuation,SpaceToken orth:upperInitial,allCaps,lowerCase,mixedCaps length stringSentenceSpliter 实现分句功能功能介绍Gazetteer 辞典 lists.def内容包括 country.lst:location:country country.lst内容包括 China Chine Chypre Colombia Colombie功能介绍PartofSpeechTagger 词性标注 也有标注错误的： Iwillstudyhardthisyear. JJ（adjective，应当为RBadverb）功能介绍SemanticTagger 就是NETransducer，命名实体识别OrthographicCoreference（Orthomatcher） TheOrthomatchermoduleaddsidentityrelationsbetweennamedentitiesfoundbythesemantictagger,inordertoperformcoreference.PronominalCoreference 将人名、代词联系起来，比如： JohnSmith…he…him…John…he…功能介绍DocumentReset Removealltheannotationsetsandtheircontents,apartfromtheonecontainingthedocumentformatanalysis(OriginalMarkups).功能介绍VerbGroupChunker Therulescoverfinite(isinvestigating),non-finite(toinvestigate),participles(investigated),andspecialverbconstructs(isgoingtoinvestigate).NounPhraseChunker Markingnounphrasesintext.功能介绍OntoTextGazetteer 与ANNIEGazetteer结果相似，但是算法不同。FlexibleGazetteer TheFlexibleGazetteerprovidesuserswiththeexibilitytochoosetheirowncustomizedinputandanexternalGazetteer.GazetteerListCollector 指定标注类型的实体插入到指定Gazetteer的相应list中 并生成统计文件（实体名$次数）功能介绍TreeTagger TheTreeTagg