资源描述:
《测量作者研究关联性:基于单词、基于主题和作者共被引方法的比较(英文)》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、MeasuringAuthorResearchRelatedness:AComparisonofWord-Based,Topic-Based,andAuthorCocitationApproachesKunLuandDietmarWolframSchoolofInformationStudies,UniversityofWisconsin-Milwaukee,P.O.Box413,Milwaukee,WI53201.E-mail:{kunlu,dwolfram}@uwm.eduRelationshipsbetweenauthorsbasedonc
2、haracteristicsrelationshipstudiedisbasedonthedataused:directcitation,ofpublishedliteraturehavebeenstudiedfordecades.cocitationanalysis,co-authorshipanalysis,bibliographicAuthorcocitationanalysisusingmappingtechniquescouplinganalysis,andco-wordanalysis(discussedbelow).hasbeenm
3、ostfrequentlyusedtostudyhowcloselytwoauthorsarethoughttobeinintellectualspacebasedonAllhavebeensuccessfullyappliedtovisualizescientifichowmembersoftheresearchcommunityco-citetheirstructureandtodescribeauthorrelatedness.Recently,moreworks.Otherapproachesexisttostudyauthorrelate
4、d-sophisticatedhybridmethods(i.e.,usingtextualcontentandnessbasedmoredirectlyonthetextoftheirpublishedcitations)havebeenappliedtothemappingofarticlesworks.Inthisstudywepresentstaticanddynamicword-(Ahlgren&Colliander,2009;Boyack&Klavans,2010;Caobasedapproachesusingvectorspacem
5、odeling,aswellasatopic-basedapproachbasedonlatentDirichletallo-&Gao,2005)andjournals(Liuetal.,2010).Tothebestofourcationformappingauthorresearchrelatedness.Vectorknowledgetheuseoftextualcontentand,morespecifically,aspacemodelingisusedtodefineanauthorspacecon-topicmodel(e.g.,Dee
6、rwester,Dumais,Furnas,Landauer,&sistingofworksbyagivenauthor.OutcomesforHarshman,1990)todeterminetherelatednessofauthorsthetwoword-basedapproachesandatopic-basedhavenotbeenstudiedyet.approachfor50prolificauthorsinlibraryandinforma-tionsciencearecomparedwithmoretraditionalautho
7、rInthisstudyweproposenewtextualfeature-basedcocitationanalysisusingmultidimensionalscalingapproachesbasedonco-occurringwordsthatapplyvectorandhierarchicalclusteranalysis.Thetwoword-basedspacemodelingtomeasuretherelatednessofauthors’approachesproducedsimilaroutcomesexceptwhere
8、research.Atopic-basedapproachusinglatentDirichlettwoauthorswerefrequ