Digital Libraries: Steps toward information finding

Digital Libraries: A model and steps toward information finding

Digital Libraries: Steps toward information finding Dr. Lillian N. CasselVillanova UniversityBut first,

and

A brief introduction to Information RetrievalResource: Christopher D. Manning, Prabhakar Raghavan and Hinrich Schtze, Introduction to Information Retrieval, Cambridge University Press. 2008. The entire book is available online, free, at http://nlp.stanford.edu/IR-book/information-retrieval-book.htmlI will use some of the slides that they provide to go with the book.Authors definitionInformation 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).Note the use of the word usually. We will see examples where the material is not documents, and not text.We have already seen that the collection may be distributed over several computers

Examples and ScalingIR is about finding a needle in a haystack finding some particular thing in a very large collection of similar things. Our examples are necessarily small, so that we can comprehend them. Do remember, that all that we say must scale to very large quantities.Just how much information?Libraries are about access to information.What sense do you have about information quantity?How fast is it growing?Are there implications for the quantity and rate of increase?How much information is there?Data summarization, trend detection anomaly detection are key technologies

YottaZettaExaPetaTeraGigaMegaKiloA BookA MovieAll books(words)All Books MultiMediaEverythingRecorded !A PhotoThe smaller scale:24 Yecto, 21 zepto, 18 atto, 15 femto, 12 pico, 9 nano, 6 micro, 3 milli Slide source Jim Gray Microsoft Research (modified)See Mike Lesk: How much information is there: http://www.lesk.com/mlesk/ksg97/ksg.htmlSee Lyman & Varian: How much informationhttp://www.sims.berkeley.edu/research/projects/how-much-info/Soon most everything will be recorded and indexedMost bytes will never be seen by humans.These require algorithms, data and knowledge representation, and knowledge of the domain8 Where does the information come from?Many sourcesCorporationsIndividualsInterest groupsNews organizationsAccumulated through crawling Basic Crawl ArchitectureWWWDNSParseContentseen?DocFPsDupURLelimURLsetURL FrontierURLfilterrobotsfiltersFetchRef: Manning Introduction to Information Retrieval10Crawler ArchitectureModules:The URL frontier (the queue of URLs still to be fetched, or fetched again)A DNS resolution module (The translation from a URL to a web server to talk to)A fetch module (use http to retrieve the page)A parsing module to extract text and links from the pageA duplicate elimination module to recognize links already seenRef: Manning Introduction to Information Retrieval11Crawling threadsWith so much space to explore, so many pages to process, a crawler will often consist of many threads, each of which cycles through the same set of steps we just saw. There may be multiple threads on one processor or threads may be distributed over many nodes in a distributed system.12PolitenessNot optional.ExplicitSpecified by the web site ownerWhat portions of the site may be crawled and what portions may not be crawledrobots.txt fileImplicitIf no restrictions are specified, still restrict how often you hit a single site.You may have many URLs from the same site. Too much traffic can interfere with the sites operation. Crawler hits are much faster than ordinary traffic could overtax the server. (Constitutes a denial of service attack) Good web crawlers do not fetch multiple pages from the same server at one time.13Robots.txtProtocol nearly as old as the webSee www.rototstxt.org/robotstxt.htmlFile: URL/robots.txtContains the access restrictionsExample:User-agent: *Disallow: /yoursite/temp/

User-agent: searchengineDisallow:

All robots (spiders/crawlers)Robot named searchengine onlyNothing disallowedSource: www.robotstxt.org/wc/norobots.html

14Another example15User-agent: *Disallow: /cgi-bin/Disallow: /tmp/Disallow: /~joe/Processing robots.txtFirst line:User-agent identifies to whom the instruction applies. * = everyone; otherwise, specific crawler nameDisallow: or Allow: provides path to exclude or include in robot access.Once the robots.txt file is fetched from a site, it does not have to be fetched every time you return to the site.Just takes time, and uses up hits on the serverCache the robots.txt file for repeated reference16Robots tagrobots.txt provides information about access to a directory.A given file may have an html meta tag that directs robot behaviorA responsible crawler will check for that tag and obey its direction.Ex:

OPTIONS: INDEX, NOINDEX, FOLLOW, NOFOLLOWSee http://www.w3.org/TR/html401/appendix/notes.html#h-B.4.1.2 and http://www.robotstxt.org/meta.html

17CrawlingPick a URL from the frontierFetch the document at the URLParse the URLExtract links from it to other docs (URLs)Check if URL has content already seenIf not, add to indicesFor each extracted URLEnsure it passes certain URL filter testsCheck if it is already in the frontier (duplicate URL elimination)

Ref: Manning Introduction to Information RetrievalWhich one?E.g., only crawl .edu, obey robots.txt, etc.18Basic Crawl ArchitectureRef: Manning Introduction to Information RetrievalWWWDNSParseContentseen?DocFPsDupURLelimURLsetURL FrontierURLfilterrobotsfiltersFetch19DNS Domain Name ServerInternet service to resolve URLs into IP addressesDistributed servers, some significant latency possibleOS implementations DNS lookup is blocking only one outstanding request at a time.SolutionsDNS cachingBatch DNS resolver collects requests and sends them out together

Ref: Manning Introduction to Information Retrieval20ParsingFetched page contains Embedded links to more pagesActual content for use in the applicationExtract the linksRelative link? Expand (normalize)Seen before? DiscardNew?Meet criteria? Append to URL frontierDoes not meet criteria? DiscardExamine content21ContentSeen before?How to tell?Finger Print, ShinglesDocuments identical, or similarIf already in the index, do not process it againRef: Manning Introduction to Information Retrieval22Distributed crawlerFor big crawls,Many processes, each doing part of the jobPossibly on different nodesGeographically distributedHow to distributeGive each node a set of hosts to crawlUse a hashing function to partition the set of hostsHow do these nodes communicate?Need to have a common index Ref: Manning Introduction to Information Retrieval23Communication between nodes

Ref: Manning Introduction to Information RetrievalWWWFetchDNSParseContentseen?URLfilterDupURLelimDocFPsURLsetURL FrontierrobotsfiltersHostsplitterTootherhostsFromotherhostsThe output of the URL filter at each node is sent to the Duplicate URL Eliminator at all nodes

24URL FrontierTwo requirementsPoliteness: do not go too often to the same siteFreshness: keep pages up to dateNews sites, for example, change frequentlyConflicts The two requirements may be directly in conflict with each other.ComplicationFetching URLs embedded in a page will yield many URLs located on the same server. Delay fetching those.Ref: Manning Introduction to Information Retrieval25Now that we have a collectionHow will we ever find the needle in the haystack? The one bit of information needed?After crawling, or other resource acquisition step, we need to create a way to query the information we haveNext step: IndexExample content: Shakespeares plays

Searching ShakespeareWhich plays of Shakespeare contain the words Brutus AND Caesar but NOT Calpurnia?See http://www.rhymezone.com/shakespeare/One could 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)

grep (general regular expression print)27Term-document incidence

1 if play contains word, 0 otherwiseBrutus AND Caesar BUT NOT CalpurniaAll the plays All the terms First approach make a matrix with terms on one axis and plays on the otherIncidence 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 AND 110111 AND 101111 = 100100.

Answer 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 i' the Capitol; Brutus killed me.

Try another oneWhat is the vector for the queryAntony and mercy

What would we do to find Antony OR mercy?

Basic assumptions about information retrievalCollection: Fixed set of documents

Goal: Retrieve documents with information that is relevant to the users information need and helps the user complete a task

The classic search modelCorpus

TASK Info Need

Query

Verbal formResults

SEARCHENGINE

QueryRefinement Ultimately, some task to perform.Some information is required in order to perform the task.The information need must be expressed in words (usually).The information need must be expressed in the form of a query that can be processed.It may be necessary to rephrase the query and try againThe classic search modelCorpus

TASK Info Need

Query

Verbal formResults

SEARCHENGINE

QueryRefinement Get rid of mice in a politically correct wayInfo about removing micewithout killing them How do I trap mice alive?

mouse trapMisconception?Mistranslation?Misformulation?Potential pitfalls between task and query resultsHow good are the results?Precision: How well do the results match the information need?

Recall: What fraction of the available correct results were retrieved?

These are the basic concepts of information retrieval evaluation.Size considerationsConsider 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.

The matrix does not work500K x 1M matrix has half-a-trillion 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.i.e. We dont need to know which documents do not have a term, only those that do.

Why?1 Million documents with 1,000 words each = 1 Billion words.37Inverted indexFor each term t, we must store a list of all documents that contain t.Identify each by a docID, a document serial numberCan we used fixed-size arrays for this?

BrutusCalpurniaCaesar12456165713212411314517317423154101What happens if we add document 14, which contains Caesar.Inverted indexWe need variable-size postings listsOn disk, a continuous run of postings is normal and bestIn memory, can use linked lists or variable length arraysSome tradeoffs in size/ease of insertion39DictionaryPostingsSorted by docID (more later on why).PostingBrutusCalpurniaCaesar12456165713212411314517323117454101TokenizerToken stream.FriendsRomansCountrymenInverted index constructionLinguistic modulesModified tokens.friendromancountrymanIndexerInverted index.friendromancountryman24213161

Documents tobe indexed.Friends, Romans, countrymen.Stop words, stemming, capitalization, cases, etc. 40Indexer steps: Token sequenceSequence of (Modified 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 2

Indexer steps: SortSort by termsAnd then docID

Core indexing stepIndexer steps: Dictionary & PostingsMultiple term entries in a single document are merged.Split into Dictionary and PostingsDoc. frequency information is added.

Where do we pay in storage?44PointersTerms and countsLists of docIDsStorageA small diversionComputer storage Processor cachesMain memoryExternal storage (hard disk, other devices)Very substantial differences in access speedsProcessor caches mostly used by the operating system for rapid access to data that will be needed soonMain memory.Limited quantities. High speed accessHard diskMuch larger quantities, speed restricted, access in fixed units (blocks) Some size examplesFrom the MacBook ProMemory8GB (two 4GB SO-DIMMs) of 1333MHz DDR3 SDRAMHard drive 500GB 7200-rpm Serial ATA hard driveThe Cloud 2 TB PogoPlugDropbox, iCloud, etc.These are big numbers, but the potential size of a significant collection is larger still. The steps taken to optimize use of storage are critical to satisfactory response time.Implications of size limitsslot 1slot 2slot 3slot 4slot 5slot 6slot 7slot 8slot 9slot 10slot 11slot 12real memory (RAM)page1page2page3page4.....................................................................b.............................................page70page71page72virtual memory (on disk)reference to page 71virtual page 71 is brought from disk into real memoryvirtual page 2 generates a page fault when referencing virtual page 71How do we process a query?Using the index we just built, examine the terms in some order, looking for the terms in the query.Query processing: ANDConsider processing the query:Brutus AND CaesarLocate Brutus in the Dictionary;Retrieve its postings.Locate Caesar in the Dictionary;Retrieve its postings.Merge the two postings:4912834248163264123581321BrutusCaesarThe mergeWalk through the two postings simultaneously, in time linear in the total number of postings entries503412824816326412358132112834248163264123581321BrutusCaesar28If the list lengths are x and y, the merge takes O(x+y) operations.What does that mean?

Crucial: postings sorted by docID.Intersecting two postings lists(a merge algorithm)51

A small exercise128342481632341358131621BrutusAntonyAssume the list of documents shown here. Show the processing of the query: Brutus AND AntonyBoolean queries: Exact 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 buildPrimary commercial retrieval tool for 3 decades. Many search systems you still use are Boolean:Email, library catalog, Mac OS X Spotlight53Ranked retrievalThus far, our queries have all been Boolean.Documents either match or dont.Good for expert users with precise understanding of their needs and the collection.Also good for applications: Applications can easily consume 1000s of results.Writing the precise query that will produce the desired result is difficult for most users.

Problem with Boolean search:feast or famineBoolean queries often result in either too few (=0) or too many (1000s) results.A query that is too broad yields hundreds of thousands of hitsA query that is too narrow may yield no hitsIt takes a lot of skill to come up with a query that produces a manageable number of hits.AND gives too few; OR gives too manyCf. our discussion of how Westlaw Boolean queries didnt actually outperform free text querying

55Ranked retrieval modelsRather than a set of documents satisfying a query expression, in ranked retrieval models, the system returns an ordering over the (top) documents in the collection with respect to a queryFree text queries: Rather than a query language of operators and expressions, the users query is just one or more words in a human languageIn principle, these are different options, but in practice, ranked retrieval models have normally been associated with free text queries and vice versa

56Feast or famine: not a problem in ranked retrievalWhen a system produces a ranked result set, large result sets are not an issueIndeed, the size of the result set is not an issueWe just show the top k ( 10) resultsWe dont overwhelm the user

Premise: the ranking algorithm worksScoring as the basis of ranked retrievalWe wish to return, in order, the documents most likely to be useful to the searcherHow can we rank-order the documents in the collection with respect to a query?Assign a score say in [0, 1] to each documentThis score measures how well document and query match.Query-document matching scoresWe need a way of assigning a score to a query/document pairLets start with a one-term queryIf the query term does not occur in the document: score should be 0The more frequent the query term in the document, the higher the score (should be)We will look at a number of alternatives for this.Take 1: Jaccard coefficientA commonly used measure of overlap of two sets A and Bjaccard(A,B) = |A B| / |A B|jaccard(A,A) = 1jaccard(A,B) = 0 if A B = 0A and B dont have to be the same size.Always assigns a number between 0 and 1.Jaccard coefficient: Scoring exampleWhat is the query-document match score that the Jaccard coefficient computes for each of the two documents below?Query: ides of marchDocument 1: caesar died in marchDocument 2: the long marchWhich do you think is probably the better match?Doc 1: 1/6 (Using all words, no stop words ignored)Doc 2: 1/561Jaccard Example doneQuery: ides of marchDocument 1: caesar died in marchDocument 2: the long marchA = {ides, of, march}B1 = {caesar, died, in, march}

B2 = {the, long, march}

Issues with Jaccard for scoringIt doesnt consider term frequency (how many times a term occurs in a document)Rare terms in a collection are more informative than frequent terms. Jaccard doesnt consider this informationWe need a more sophisticated way of normalizing for lengthLater, well use . . . instead of |A B|/|A B| (Jaccard) for length normalization.

Recall : Binary term-document incidence matrix

Each document is represented by a binary vector {0,1}|V|Term-document count matricesConsider the number of occurrences of a term in a document: Each document is a count vector in v: a column below

Bag of words modelVector representation doesnt consider the ordering of words in a documentJohn is quicker than Mary and Mary is quicker than John have the same vectorsThis is called the bag of words model.There are ways of recovering positional information.For now: bag of words modelTerm frequency - tfThe term frequency tft,d of term t in document d is defined as the number of times that t occurs in d.We want to use tf when computing query-document match scores. But how?Raw term frequency is not what we want:A document with 10 occurrences of the term is more relevant than a document with 1 occurrence of the term.But not 10 times more relevant.Relevance does not increase proportionally with term frequency.NB: frequency = count in IRLog-frequency weightingThe log frequency weight of term t in document d is

0 0, 1 1, 2 1.3, 10 2, 1000 4, etc.Score for a document-query pair: sum over terms t in both q and d:

The score is 0 if none of the query terms is present in the document.

score

Document frequencyRare terms are more informative than frequent termsConsider stop wordsConsider a term in the query that is rare in the collection (e.g., arachnocentric)A document containing this term is very likely to be relevant to the query arachnocentric We want a high weight for rare terms like arachnocentric, even if the term does not appear many times in the document.Document frequency, continuedFrequent terms are less informative than rare termsConsider a query term that is frequent in the collection (e.g., high, increase, line)A document containing such a term is more likely to be relevant than a document that doesntBut its not a sure indicator of relevance. For frequent terms, we want high positive weights for words like high, increase, and lineBut lower weights than for rare terms.We will use document frequency (df) to capture this.

idf weightdft is the document frequency of t: the number of documents that contain tdft is an inverse measure of the informativeness of tdft N (the number of documents)We define the idf (inverse document frequency) of t by

We use log (N/dft) instead of N/dft to dampen the effect of idf.

Will turn out the base of the log is immaterial.idf example, suppose N = 1 milliontermdftidftcalpurnia1animal100sunday1,000fly10,000under100,000the1,000,000There is one idf value for each term t in a collection.

6432106 4 3 2 1 072Collection vs. Document frequencyThe collection frequency of t is the number of occurrences of t in the collection, counting multiple occurrences.Example:

Which word is a better search term (and should get a higher weight)?WordCollection frequencyDocument frequencyinsurance104403997try104228760Why do you get these numbers?Suggests df is better.

73tf-idf weightingThe tf-idf weight of a term is the product of its tf weight and its idf weight.

Best known weighting scheme in information retrievalNote: the - in tf-idf is a hyphen, not a minus sign!Alternative names: tf.idf, tf x idfIncreases with the number of occurrences within a documentIncreases with the rarity of the term in the collection

Final ranking of documents for a query75

Binary count weight matrix

Each document is now represented by a real-valued vector of tf-idf weights R|V|Documents as vectorsSo we have a |V|-dimensional vector spaceTerms are axes of the spaceDocuments are points or vectors in this spaceVery high-dimensional: tens of millions of dimensions when you apply this to a web search engineThese are very sparse vectors - most entries are zero.Queries as vectorsKey idea 1: Do the same for queries: represent them as vectors in the spaceKey idea 2: Rank documents according to their proximity to the query in this spaceproximity = similarity of vectorsproximity inverse of distanceRecall: We do this because we want to get away from the youre-either-in-or-out Boolean model.Instead: rank more relevant documents higher than less relevant documentsFormalizing vector space proximityFirst cut: distance between two points( = distance between the end points of the two vectors)Euclidean distance?Euclidean distance is a bad idea . . .. . . because Euclidean distance is large for vectors of different lengths.Why distance is a bad idea

The Euclidean distance between qand d2 is large even though thedistribution of terms in the query q and the distribution ofterms in the document d2 arevery similar.We are showing only two-dimensions here, for the sake of drawing. Vectors are usually much higher dimension a dimension for each term of interest.Use angle instead of distanceThought experiment: take a document d and append it to itself. Call this document d.Semantically d and d have the same contentThe Euclidean distance between the two documents can be quite largeThe angle between the two documents is 0, corresponding to maximal similarity.

Key idea: Rank documents according to angle with query.From angles to cosinesThe following two notions are equivalent.Rank documents in decreasing order of the angle between query and documentRank documents in increasing order of cosine(query,document)Cosine is a monotonically decreasing function for the interval [0o, 180o]From angles to cosinesBut how and why should we be computing cosines?

Cosine similarity illustrated84

Length normalizationA vector can be (length-) normalized by dividing each of its components by its length for this we use the L2 norm:

Dividing a vector by its L2 norm makes it a unit (length) vector (on surface of unit hypersphere)Effect on the two documents d and d (d appended to itself) from earlier slide: they have identical vectors after length-normalization.Long and short documents now have comparable weights

cosine(query,document)

Dot productUnit vectorsqi is the tf-idf weight of term i in the querydi is the tf-idf weight of term i in the document

cos(q,d) is the cosine similarity of q and d or,equivalently, the cosine of the angle between q and d.Remember: tf gives significance to terms that appear frequently; idf gives significance to terms that are unusual over all the documentsSee Law of Cosines (Cosine Rule) wikipedia page86Cosine for length-normalized vectorsFor length-normalized vectors, cosine similarity is simply the dot product (or scalar product):

for q, d length-normalized.

87

Cosine similarity amongst 3 documentstermSaSPaPWHaffection11558 20jealous10711gossip206wuthering0038How similar arethe novelsSaS: Sense andSensibilityPaP: Pride andPrejudice, andWH: WutheringHeights?Term frequencies (counts)Note: To simplify this example, we dont do idf weighting.3 documents example contd.Log frequency weightingtermSaSPaPWHaffection3.062.762.30jealous2.001.852.04gossip1.3001.78wuthering002.58After length normalizationtermSaSPaPWHaffection0.7890.8320.524jealous0.5150.5550.465gossip0.33500.405wuthering000.588cos(SaS,PaP) 0.789 0.832 + 0.515 0.555 + 0.335 0.0 + 0.0 0.0 0.94cos(SaS,WH) 0.79cos(PaP,WH) 0.69Why do we have cos(SaS,PaP) > cos(SAS,WH)?tf-idf weighting has many variants

Columns headed n are acronyms for weight schemes.Why is the base of the log in idf immaterial?n default is just term frequencyltc is best known form of weighting90Weighting may differ in queries vs documentsMany search engines allow for different weightings for queries vs. documentsSMART Notation: denotes the combination in use in an engine, with the notation ddd.qqq, using the of the previous table.A very standard weighting scheme is: lnc.ltcDocument: logarithmic tf (l as first character), no idf and cosine normalizationQuery: logarithmic tf (l in leftmost column), idf (t in second column), no cosine normalization Leaving off idf weighting on documents is good for both efficiency and system effectiveness reasons.91tf-idf example: lnc.ltcTermQueryDocumentProdtf-rawtf-wtdfidfwtnlizetf-rawtf-wtwtnlizeauto0050002.3 001110.520best11500001.31.30.3400000car1 1100002.02.00.521110.520.27insurance1110003.03.00.7821.31.30.680.53Document: car insurance auto insuranceQuery: best car insuranceExercise: what is N, the number of docs?Score = 0+0+0.27+0.53 = 0.8Doc length term =

Idf = log10(N/df); 3 = log10(N/1000); 10^3=N/1000; 1000000=NWorks with idf=2 also: 2=log10(N/10000); 100*10000=N N=1,000,00092Summary vector space rankingRepresent the query as a weighted tf-idf vectorRepresent each document as a weighted tf-idf vectorCompute the cosine similarity score for the query vector and each document vectorRank documents with respect to the query by scoreReturn the top K (e.g., K = 10) to the userThe web and its challengesUnusual and diverse documentsUnusual and diverse users, queries, information needsBeyond terms, exploit ideas from social networkslink analysis, clickstreams How do search engines work? And how can we make them better?See http://video.google.com/videoplay?docid=-1243280683715323550&hl=en#Marissa Mayer of Google on how a search happens at Google. 9494ReferencesChristopher D. Manning, Prabhakar Raghavan and Hinrich Schtze, Introduction to Information Retrieval, Cambridge University Press. 2008. Book available online at http://nlp.stanford.edu/IR-book/information-retrieval-book.htmlMany of these slides are taken directly from the authors slides from the first chapter of the book.Slides provided by the authorPaging figure from Vittore Casarosahttp://www.miislita.com/information-retrieval-tutorial/cosine-similarity-tutorial.html -- Term weighting and cosine similarity tutorial

ShakepeareAntony and CleopatraJulius CaesarThe TempestHamletOthelloMacbethAntony110001Brutus110100Caesar110111Calpurnia010000Cleopatra100000mercy101111worser101110

Sheet1TermdocIDI1did1enact1julius1caesar1I1was1killed1i'1the1capitol1brutus1killed1me1so2let2it2be2with2caesar2the2noble2brutus2hath2told2you2caesar2was2ambitious2

Sheet1TermdocIDambitious2be2brutus1brutus2capitol1caesar1caesar2caesar2did1enact1hath1I1I1i'1it2julius1killed1killed1let2me1noble2so2the1the2told2you2was1was2with2

Sheet1TermdocIDI1did1enact1julius1caesar1I1was1killed1i'1the1capitol1brutus1killed1me1so2let2it2be2with2caesar2the2noble2brutus2hath2told2you2caesar2was2ambitious2

Sheet1TermdocIDambitious2be2brutus1brutus2capitol1caesar1caesar2caesar2did1enact1hath1I1I1i'1it2julius1killed1killed1let2me1noble2so2the1the2told2you2was1was2with2



ShakepeareAntony and CleopatraJulius CaesarThe TempestHamletOthelloMacbethAntony5.253.180000.35Brutus1.216.10100Caesar8.592.5401.510.250Calpurnia01.540000Cleopatra2.8500000mercy1.5101.90.125.250.88worser1.3700.114.150.251.95