资源描述：
《and computer typesettingnew》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、MasarykUniversityinBrnoFacultyofInformatics
!"#$%GHI &'EF(*),+@ABDC-/.103254687:9/;=CompetingPatternsinLanguageEngineeringandComputerTypesettingPetrSojkaPh.D.DissertationSupervisor:KarelPalaBrno,January2005CompetingPatternsinLanguageEngineer
2、ingandComputerTypesettingADissertationPresentedtotheFacultyofInformaticsofMasarykUniversityinBrnoinPartialFullllmentoftheRequirementsforthePh.D.DegreebyPetrSojkaJanuary2005AbstractThegoalofthisdissertationistoexploremodels,methodsandmethodologiesformachinelearningofthecom
3、pactandeffectivestorageofempiricaldataintheareasoflanguageengineeringandcomputertypesetting,withafocusonthemassiveexceptionhandling.Researchhasfocusedonthepattern-drivenapproach.Thewholemethodologyofsocalledcompetingpatternscapableofhandlingexceptionstobefoundsowidelyinnat
4、urallanguagedataandcomputertypesetting,isfurtherdeveloped.Competingpatternscanstorecontextdependentinforma-tionandcanbelearntfromdata,orwrittenbyexperts,orcombinedtogether.Intherstpartofthethesis,thetheoryofcompetingpatternsisbuilt;competingpatternsaredened,cornerstoneso
5、fmethodologybasedonstratiedsampling,bootstrappingandproblemmodelingbycompetingpatternsaredescribed.Segmentationproblems(hyphenation)andproblemsofdisambiguationoftaggeddataincorpuslinguisticsareusedasexampleswhendevelopingformalmodelofthecompetingpatternsmethod.Thesecondpa
6、rtconsistofaseriesofsevenpublishedpapersthatde-scribeproblemsaddressedbytheproposedmethods:applicationsofcompet-ingpatternsandrelatedlearningmethodsinareasofhyphenation,hyphen-ationofcompoundwordsand,forexample,thesegmentationofThaitexts.KeyWordscompetingpatternscontext-s
7、ensitivepatternsmachinelearningnaturallanguageengineeringhyphenationsegmentationThaitextsegmentationdisambiguationformalconceptanalysispartofspeechtaggingiiiDeclarationIdeclarethatthisdissertationthesisistheresultofmyworkcarriedoutattheFacultyofInformatics,MasarykU
8、niversityinBrno,CzechRepublic.Theworkdescribedhere,istothebestofmyknowledge,allmyown,exce
