剑桥Information retrieval(信息检索)课件3.pdf

《剑桥Information retrieval(信息检索)课件3.pdf》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、Onlineedition(c)2009CambridgeUPDRAFT!©April1,2009CambridgeUniversityPress.Feedbackwelcome.493DictionariesandtolerantretrievalInChapters1and2wedevelopedtheideasunderlyinginvertedindexesforhandlingBooleanandproximityqueries.Here,wedeveloptechniquesthatarerobusttotypographicalerro

2、rsinthequery,aswellasalternativespellings.InSection3.1wedevelopdatastructuresthathelpthesearchfortermsinthevocabularyinaninvertedindex.InSection3.2westudyWILDCARDQUERYtheideaofawildcardquery:aquerysuchas*a*e*i*o*u*,whichseeksdoc-umentscontaininganytermthatincludesallthefivevowel

3、sinsequence.The*symbolindicatesany(possiblyempty)stringofcharacters.Usersposesuchqueriestoasearchenginewhentheyareuncertainabouthowtospellaqueryterm,orseekdocumentscontainingvariantsofaqueryterm;forin-stance,thequeryautomat*wouldseekdocumentscontaininganyofthetermsautomatic,aut

4、omationandautomated.Wethenturntootherformsofimpreciselyposedqueries,focusingonspellingerrorsinSection3.3.Usersmakespellingerrorseitherbyaccident,orbecausethetermtheyaresearchingfor(e.g.,Herman)hasnounambiguousspellinginthecollection.Wedetailanumberoftechniquesforcorrectingspell

5、ingerrorsinqueries,onetermatatimeaswellasforanentirestringofqueryterms.Finally,inSection3.4westudyamethodforseekingvo-cabularytermsthatarephoneticallyclosetothequeryterm(s).ThiscanbeespeciallyusefulincasesliketheHermanexample,wheretheusermaynotknowhowapropernameisspelledindocum

6、entsinthecollection.Becausewewilldevelopmanyvariantsofinvertedindexesinthischapter,wewillusesometimesthephrasestandardinvertedindextomeantheinvertedindexdevelopedinChapters1and2,inwhicheachvocabularytermhasapostingslistwiththedocumentsinthecollection.3.1Searchstructuresfordicti

7、onariesGivenaninvertedindexandaquery,ourfirsttaskistodeterminewhethereachquerytermexistsinthevocabularyandifso,identifythepointertotheOnlineedition(c)2009CambridgeUP503Dictionariesandtolerantretrievalcorrespondingpostings.Thisvocabularylookupoperationusesaclassicaldatastructurec

8、alledthedictionaryandhastwobroadclassesofsolutions:has