Modern Information Retrieval A Brief Overview.pdf

《Modern Information Retrieval A Brief Overview.pdf》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、ModernInformationRetrieval:ABriefOverviewAmitSinghalGoogle,Inc.singhal@google.comAbstractForthousandsofyearspeoplehaverealizedtheimportanceofarchivingandfindinginformation.Withtheadventofcomputers,itbecamepossibletostorelargeamountsofinformation;andfindingusefulinformationfromsuchcol

2、lectionsbecameanecessity.ThefieldofInformationRetrieval(IR)wasborninthe1950soutofthisnecessity.Overthelastfortyyears,thefieldhasmaturedconsiderably.SeveralIRsystemsareusedonaneverydaybasisbyawidevarietyofusers.ThisarticleisabriefoverviewofthekeyadvancesinthefieldofInformationRetrieval

3、,andadescriptionofwherethestate-of-the-artisatinthefield.1BriefHistoryThepracticeofarchivingwritteninformationcanbetracedbacktoaround3000BC,whentheSumeriansdesignatedspecialareastostoreclaytabletswithcuneiforminscriptions.EventhentheSumeriansrealizedthatproperorganizationandaccessto

4、thearchiveswascriticalforefficientuseofinformation.Theydevelopedspecialclassificationstoidentifyeverytabletanditscontent.(Seehttp://www.libraries.grforawonderfulhistoricalperspectiveonmodernlibraries.)Theneedtostoreandretrievewritteninformationbecameincreasinglyimportantovercenturies

5、,especiallywithinventionslikepaperandtheprintingpress.Soonaftercomputerswereinvented,peoplerealizedthattheycouldbeusedforstoringandmechanicallyretrievinglargeamountsofinformation.In1945VannevarBushpublishedagroundbreakingarticletitled“AsWeMayThink”thatgavebirthtotheideaofautomatica

6、ccesstolargeamountsofstoredknowledge.[5]Inthe1950s,thisideamaterializedintomoreconcretedescriptionsofhowarchivesoftextcouldbesearchedautomatically.Severalworksemergedinthemid1950sthatelaborateduponthebasicideaofsearchingtextwithacomputer.OneofthemostinfluentialmethodswasdescribedbyH

7、.P.Luhnin1957,inwhich(putsimply)heproposedusingwordsasindexingunitsfordocumentsandmeasuringwordoverlapasacriterionforretrieval.[17]Severalkeydevelopmentsinthefieldhappenedinthe1960s.MostnotablewerethedevelopmentoftheSMARTsystembyGerardSaltonandhisstudents,firstatHarvardUniversityandl

8、ateratCornellUniver-sity;[