欢迎来到天天文库
浏览记录
ID:40103322
大小:1.77 MB
页数:243页
时间:2019-07-21
《Text Mining - Predictive Methods for Analyzing Unstructured Information,Weiss ,2005》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、TextMiningSholomM.WeissNitinIndurkhyaTongZhangFredJ.DamerauTextMiningPredictiveMethodsforAnalyzingUnstructuredInformation~SpringerSholomM.WeissNitinIndurkhyaIBMResearchSchoolofComputerScienceandEngineeringTJWatsonLabsUniversityofNewSouthWalesYorktownHeights)NY1
2、0598Sydney)NSW2052USAAustraliasholom@us.ibm.comnitin@data-miner.comTongZhangFredJ.DamerauIBMResearchIBMResearchTJWatsonLabsTJWatsonLabsYorktownHeights)NY10598YorktownHeights)NY10598USAUSAtongz@us.ibm.comdamerau@sbcglobal.netISBN0-387-95433-3Printedonacid-freepa
3、per.©2005SpringerScience+BusinessMedia)Inc.Allrightsreserved.Thisworkmaynotbetranslatedorcopiedinwholeorinpartwithoutthewrittenpermissionofthepublisher(SpringerScience+BusinessMedia)Inc.,233SpringStreet)NewYork)NY10013)USA))exceptforbriefexcerptsincon-nectionwi
4、threviewsorscholarlyanalysis.Useinconnectionwithanyformofinfor-mationstorageandretrieval)electronicadaptation)computersoftware)orbysimilarordissimilarmethodologynowknownorhereafterdevelopedisforbidden.Theuseinthispublicationoftradenames)trademarks)servicemarks)
5、andsimilarterms)eveniftheyarenotidentifiedassuch)isnottobetakenasanexpressionofopinionastowhetherornottheyaresubjecttoproprietaryrights.PrintedintheUnitedStatesofAmerica.(MP)987654321SPIN10864579springeronline.comPrefaceDataminingisamaturetechnology.Thepredicti
6、onproblem,lookingforpredictivepatternsindata,hasbeenwidelystudied.Strongmeth-odsareavailabletothepractitioner.Thesemethodsprocessstructurednumericalinformation,whereuniformmeasurementsaretakenoverasampleofdata.Textisoftendescribedasunstructuredinformation.So,it
7、wouldseem,textandnumericaldataaredifferent,requiringdifferentmethods.Orarethey?Inourview,apredictionproblemcanbesolvedbythesamemethods,whetherthedataarestructurednu-mericalmeasurementsorunstructuredtext.Textanddocumentscanbetransformedintomeasuredvalues,suchast
8、hepresenceorabsenceofwords,andthesamemethodsthathaveprovensuccessfulforpredic-tivedataminingcanbeappliedtotext.Yet,therearekeydifferences.Evaluationtechniquesmustbeadaptedto
此文档下载收益归作者所有