信息检索导论-第一章-布尔检索(英文).ppt

《信息检索导论-第一章-布尔检索(英文).ppt》由会员分享，可在线阅读，更多相关《信息检索导论-第一章-布尔检索(英文).ppt（57页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、Introduction toInformation RetrievalIntroducing Information Retrieval and Web SearchInformation RetrievalInformation Retrieval(IR)is finding material(usually documents)of an unstructured nature(usually text)that satisfies an information need from within large collections(usually stored on computers)

2、.These days we frequently think first of web search,but there are many other cases:E-mail searchSearching your laptopCorporate knowledge basesLegal information retrieval2Unstructured(text)vs.structured(database)data in the mid-nineties3Unstructured(text)vs.structured(database)data today4Basic assump

3、tions of Information RetrievalCollection:A set of documentsAssume it is a static collection for the momentGoal:Retrieve documents with information that is relevant to the users information need and helps the user complete a task5Sec.1.1how trap mice aliveThe classic search modelCollectionUser task I

4、nfo needQueryResultsSearchengineQueryrefinement Get rid of mice in a politically correct wayInfo about removing micewithout killing them Misconception?Misformulation?SearchHow good are the retrieved docs?Precision:Fraction of retrieved docs that are relevant to the users information needRecall:Fract

5、ion of relevant docs in collection that are retrievedMore precise definitions and measurements to follow later7Sec.1.1Introduction toInformation RetrievalTerm-document incidence matricesUnstructured data in 1620Which plays of Shakespeare contain the words Brutus AND Caesar but NOT Calpurnia?One coul

6、d grep all of Shakespeares plays for Brutus and Caesar,then strip out lines containing Calpurnia?Why is that not the answer?Slow(for large corpora)NOT Calpurnia is non-trivialOther operations(e.g.,find the word Romans near countrymen)not feasibleRanked retrieval(best documents to return)Later lectur

7、es9Sec.1.1Term-document incidence matrices1 if play contains word,0 otherwiseBrutus AND Caesar BUT NOT CalpurniaSec.1.1Incidence vectorsSo we have a 0/1 vector for each term.To answer query:take the vectors for Brutus,Caesar and Calpurnia(complemented)bitwise AND.110100 AND110111 AND101111=10010011S

8、ec.1.1Answers to queryAntony and Cleopatra,Act III,Scene iiAgrippa Aside to DOMITIUS ENOBARBUS:Why,Enobarbus,When Antony found Julius Caesar dead,He cried almost to roaring;and he wept When at Philippi he found Brutus slain.Hamlet,Act III,Scene iiLord Polonius:I did enact Julius Caesar I was killed

9、i the Capitol;Brutus killed me.12Sec.1.1Bigger collectionsConsider N=1 million documents,each with about 1000 words.Avg 6 bytes/word including spaces/punctuation 6GB of data in the documents.Say there are M=500K distinct terms among these.13Sec.1.1Cant build the matrix500K x 1M matrix has half-a-tri

10、llion 0s and 1s.But it has no more than one billion 1s.matrix is extremely sparse.Whats a better representation?We only record the 1 positions.14Why?Sec.1.1Introduction toInformation RetrievalThe Inverted IndexThe key data structure underlying modern IRInverted indexFor each term t,we must store a l

11、ist of all documents that contain t.Identify each doc by a docID,a document serial numberCan we used fixed-size arrays for this?16What happens if the word Caesar is added to document 14?Sec.1.2BrutusCalpurniaCaesar1245616 57 1321241131 4517323117454101Inverted indexWe need variable-size postings lis

12、tsOn disk,a continuous run of postings is normal and bestIn memory,can use linked lists or variable length arraysSome tradeoffs in size/ease of insertion17DictionaryPostingsSorted by docID(more later on why).PostingSec.1.2BrutusCalpurniaCaesar1245616 57 1321241131 4517323117454101TokenizerToken stre

13、amFriendsRomansCountrymenInverted index constructionLinguistic modulesModified tokensfriendromancountrymanIndexerInverted indexfriendromancountryman24213161Documents tobe indexedFriends,Romans,countrymen.Sec.1.2Initial stages of text processingTokenizationCut character sequence into word tokensDeal

14、with“Johns”,a state-of-the-art solutionNormalizationMap text and query term to same formYou want U.S.A.and USA to matchStemmingWe may wish different forms of a root to matchauthorize,authorizationStop wordsWe may omit very common words(or not)the,a,to,ofIndexer steps:Token sequenceSequence of(Modifi

15、ed token,Document ID)pairs.I did enact JuliusCaesar I was killed i the Capitol;Brutus killed me.Doc 1So let it be withCaesar.The nobleBrutus hath told youCaesar was ambitiousDoc 2Sec.1.2Indexer steps:SortSort by termsAnd then docID Core indexing stepSec.1.2Indexer steps:Dictionary&PostingsMultiple t

16、erm entries in a single document are merged.Split into Dictionary and PostingsDoc.frequency information is added.Why frequency?Will discuss later.Sec.1.2Where do we pay in storage?23PointersTerms and countsIR system implementationHow do we index efficiently?How much storage do we need?Sec.1.2Lists o

17、f docIDsIntroduction toInformation RetrievalQuery processing with an inverted indexThe index we just builtHow do we process a query?Later-what kinds of queries can we process?25Our focusSec.1.3Query processing:ANDConsider processing the query:Brutus AND CaesarLocate Brutus in the Dictionary;Retrieve

18、 its postings.Locate Caesar in the Dictionary;Retrieve its postings.“Merge”the two postings(intersect the document sets):2612834248163264123581321BrutusBrutusCaesarCaesarSec.1.3The mergeWalk through the two postings simultaneously,in time linear in the total number of postings entries273412824816326

19、4123581321BrutusBrutusCaesarCaesarIf the list lengths are x and y,the merge takes O(x+y)operations.Crucial:postings sorted by docID.Sec.1.3Intersecting two postings lists(a“merge”algorithm)28Introduction toInformation RetrievalThe Boolean Retrieval Model&Extended Boolean ModelsBoolean queries:Exact

20、matchThe Boolean retrieval model is being able to ask a query that is a Boolean expression:Boolean Queries are queries using AND,OR and NOT to join query termsViews each document as a set of wordsIs precise:document matches condition or not.Perhaps the simplest model to build an IR system onPrimary

21、commercial retrieval tool for 3 decades.Many search systems you still use are Boolean:Email,library catalog,Mac OS X Spotlight30Sec.1.3Example:WestLaw http:/ commercial(paying subscribers)legal search service(started 1975;ranking added 1992;new federated search added 2010)Tens of terabytes of data;7

22、00,000 usersMajority of users still use boolean queriesExample query:What is the statute of limitations in cases involving the federal tort claims act?LIMIT!/3 STATUTE ACTION/S FEDERAL/2 TORT/3 CLAIM/3=within 3 words,/S=in same sentence31Sec.1.4Example:WestLaw http:/ example query:Requirements for d

23、isabled people to be able to access a workplacedisabl!/p access!/s work-site work-place(employment/3 placeNote that SPACE is disjunction,not conjunction!Long,precise queries;proximity operators;incrementally developed;not like web searchMany professional searchers still like Boolean searchYou know e

24、xactly what you are gettingBut that doesnt mean it actually works better.Sec.1.4Boolean queries:More general mergesExercise:Adapt the merge for the queries:Brutus AND NOT CaesarBrutus OR NOT CaesarCan we still run through the merge in time O(x+y)?What can we achieve?33Sec.1.3MergingWhat about an arb

25、itrary Boolean formula?(Brutus OR Caesar)AND NOT(Antony OR Cleopatra)Can we always merge in“linear”time?Linear in what?Can we do better?34Sec.1.3Query optimizationWhat is the best order for query processing?Consider a query that is an AND of n terms.For each of the n terms,get its postings,then AND

26、them together.BrutusCaesarCalpurnia123581621 342481632 6412813 16Query:Brutus AND Calpurnia AND Caesar35Sec.1.3Query optimization exampleProcess in order of increasing freq:start with smallest set,then keep cutting further.36This is why we keptdocument freq.in dictionaryExecute the query as(Calpurni

27、a AND Brutus)AND Caesar.Sec.1.3BrutusCaesarCalpurnia123581621 342481632 6412813 16More general optimizatione.g.,(madding OR crowd)AND(ignoble OR strife)Get doc.freq.s for all terms.Estimate the size of each OR by the sum of its doc.freq.s(conservative).Process in increasing order of OR sizes.37Sec.1

28、.3ExerciseRecommend a query processing order forWhich two terms should we process first?38(tangerine OR trees)AND(marmalade OR skies)AND(kaleidoscope OR eyes)Query processing exercisesExercise:If the query is friends AND romans AND(NOT countrymen),how could we use the freq of countrymen?Exercise:Ext

29、end the merge to an arbitrary Boolean query.Can we always guarantee execution in time linear in the total postings size?Hint:Begin with the case of a Boolean formula query:in this,each query term appears only once in the query.39ExerciseTry the search feature at http:/ down five search features you

30、think it could do better40Introduction toInformation RetrievalPhrase queries and positional indexesPhrase queriesWe want to be able to answer queries such as“stanford university”as a phraseThus the sentence“I went to university at Stanford”is not a match.The concept of phrase queries has proven easi

31、ly understood by users;one of the few“advanced search”ideas that worksMany more queries are implicit phrase queriesFor this,it no longer suffices to store only entriesSec.2.4A first attempt:Biword indexesIndex every consecutive pair of terms in the text as a phraseFor example the text“Friends,Romans

32、,Countrymen”would generate the biwordsfriends romansromans countrymenEach of these biwords is now a dictionary termTwo-word phrase query-processing is now immediate.Sec.2.4.1Longer phrase queriesLonger phrases can be processed by breaking them downstanford university palo alto can be broken into the

33、 Boolean query on biwords:stanford university AND university palo AND palo altoWithout the docs,we cannot verify that the docs matching the above Boolean query do contain the phrase.Can have false positives!Sec.2.4.1Issues for biword indexesFalse positives,as noted beforeIndex blowup due to bigger d

34、ictionaryInfeasible for more than biwords,big even for themBiword indexes are not the standard solution(for all biwords)but can be part of a compound strategySec.2.4.1Solution 2:Positional indexesIn the postings,store,for each term the position(s)in which tokens of it appear:Sec.2.4.2Positional inde

35、x exampleFor phrase queries,we use a merge algorithm recursively at the document levelBut we now need to deal with more than just equalityWhich of docs 1,2,4,5could contain“to beor not to be”?Sec.2.4.2Processing a phrase queryExtract inverted index entries for each distinct term:to,be,or,not.Merge t

36、heir doc:position lists to enumerate all positions with“to be or not to be”.to:2:1,17,74,222,551;4:8,16,190,429,433;7:13,23,191;.be:1:17,19;4:17,191,291,430,434;5:14,19,101;.Same general method for proximity searchesSec.2.4.2Proximity queriesLIMIT!/3 STATUTE/3 FEDERAL/2 TORT Again,here,/k means“with

37、in k words of”.Clearly,positional indexes can be used for such queries;biword indexes cannot.Exercise:Adapt the linear merge of postings to handle proximity queries.Can you make it work for any value of k?This is a little tricky to do correctly and efficientlySee Figure 2.12 of IIRSec.2.4.2Positiona

38、l index sizeA positional index expands postings storage substantiallyEven though indices can be compressedNevertheless,a positional index is now standardly used because of the power and usefulness of phrase and proximity queries whether used explicitly or implicitly in a ranking retrieval system.Sec

39、.2.4.2Positional index sizeNeed an entry for each occurrence,not just once per documentIndex size depends on average document sizeAverage web page has 1000 termsSEC filings,books,even some epic poems easily 100,000 termsConsider a term with frequency 0.1%Why?1001100,000111000Positional postingsPosti

40、ngsDocument sizeSec.2.4.2Rules of thumbA positional index is 24 as large as a non-positional indexPositional index size 3550%of volume of original textCaveat:all of this holds for“English-like”languagesSec.2.4.2Combination schemesThese two approaches can be profitably combinedFor particular phrases(

41、“Michael Jackson”,“Britney Spears”)it is inefficient to keep on merging positional postings listsEven more so for phrases like“The Who”Williams et al.(2004)evaluate a more sophisticated mixed indexing schemeA typical web query mixture was executed in of the time of using just a positional indexIt re

42、quired 26%more space than having a positional index aloneSec.2.4.3Introduction toInformation RetrievalStructured vs.Unstructured DataIR vs.databases:Structured vs unstructured dataStructured data tends to refer to information in“tables”55EmployeeManagerSalarySmithJones50000ChangSmith6000050000IvySmi

43、thTypically allows numerical range and exact match(for text)queries,e.g.,Salary 60000 AND Manager=Smith.Unstructured dataTypically refers to free textAllowsKeyword queries including operatorsMore sophisticated“concept”queries e.g.,find all web pages dealing with drug abuseClassic model for searching

44、 text documents56Semi-structured dataIn fact almost no data is“unstructured”E.g.,this slide has distinctly identified zones such as the Title and Bullets to say nothing of linguistic structureFacilitates“semi-structured”search such asTitle contains data AND Bullets contain searchOr evenTitle is about Object Oriented Programming AND Author something like stro*rup where*is the wild-card operator57