Information Retrieval

Lecture 1Introduction to Information Retrieval (Manning et al. 2007)

Chapter 1

For the MSc Computer Science Programme

Dell ZhangBirkbeck, University of London

What is IR?

IR is about search engines.

Search Engine

Unstructured Data (Text)Keyword Queries

Database System

Structured Data (Tables)SQL Queries

VS

Structured data has been the big commercial success (e.g., Oracle) but unstructured data is now becoming dominant in a large and increasing range of activities.

Search is Cool

Text everywhere books, documents (ms-word, pdf, etc.), articles

(journal, magazine, newspaper, etc.), Web pages, emails, SMS, chat, …

Much more than text music, photos, videos, …

Much more than the Web personal enterprise, institutional and domain-specific

Search is Cool

Not only search/finding, but also organization and mining clustering classification ……

For example, given thousaunds of CVs, …

http://commons.wikimedia.org/wiki/Image:Bookshelf.jpg

Search is Hot

Most people’s means of information access In the 1990s: other people Nowadays: Web search

“92% of Internet users say the Internet is a good place to go for getting everyday information” – (2004 Pew Internet Survey)

The Search Engine War Google, Yahoo, Microsoft, …

Video: How Big Can You Think?

Data: Unstructured vs. Structured

0

20

40

60

80

100

120

140

160

Data volume Market Cap

UnstructuredStructured

in 1996

Data: Unstructured vs. Structured

0

20

40

60

80

100

120

140

160

Data volume Market Cap

UnstructuredStructured

in 2006

Free Textbook Online

C. D. Manning, P. Raghavan and H. Schütze, Introduction to Information Retrieval, Cambridge University Press. 2008. http://www-csli.stanford.edu/~schuetze/

information-retrieval-book.html

Head of Yahoo! Research

Don’t remember. Search!

Another reason for you to come to this module.

A Simple Search Engine

http://www.rhymezone.com/shakespeare/ Which plays of Shakespeare contain the words

Brutus AND Caesar but NOT Calpurnia? One could grep all of Shakespeare’s plays for

Brutus and Caesar, then strip out lines containing Calpurnia? Slow (for large corpora) NOT Calpurnia is non-trivial Other operations (e.g., find the word Romans near

countrymen) not feasible

“to be, or not to be”

How do you search in a book?

Boolean Queries

Boolean Queries are queries using AND, OR and NOT together with query terms. Each document is viewed as a set of words. Precise: a document matches condition or not. Primary commercial retrieval tool for 3 decades. Professional searchers still like Boolean queries -

you know exactly what you’re getting. For example, http://www.westlaw.com/ .

Antony and Cleopatra Julius Caesar The Tempest Hamlet Othello Macbeth

Antony 1 1 0 0 0 1

Brutus 1 1 0 1 0 0

Caesar 1 1 0 1 1 1

Calpurnia 0 1 0 0 0 0

Cleopatra 1 0 0 0 0 0

mercy 1 0 1 1 1 1

worser 1 0 1 1 1 0

Term-Document Incidence

1 if play contains word, 0 otherwiseBrutus AND Caesar but NOT

Calpurnia

Incidence Vectors

So we have a 0/1 vector for each term. To answer the query

take the vectors for Brutus, Caesar and Calpurnia (complemented) bitwise AND.

110100 & 110111 & 101111 = 100100

Query Results

Antony and Cleopatra, Act III, Scene ii……Agrippa [Aside to DOMITIUS ENOBARBUS]: Why, Enobarbus,

When Antony found Julius Caesar dead,He cried almost to roaring; and he weptWhen at Philippi he found Brutus slain.

……

Hamlet, Act III, Scene ii…….Lord Polonius: I did enact Julius Caesar I was killed i' the Capitol; Brutus killed me.……

Bigger Corpora

Consider n = 1M documents each with about 1K terms. Avg 6 bytes/term incl spaces/punctuation

6GB of data in the documents. Say there are m = 500K distinct terms among

these.

Bigger Corpora

The 500K x 1M matrix has half-a-trillion 0’s and 1’s, but no more than one billion 1’s.

The matrix is extremely sparse.

Can’t build the matrix straightforwardly. What’s a better representation?

Only record the 1 positions.

Why?

Inverted Index

For each term T, we must store a list of all documents that contain T.

Do we use an array or a list for this?

Brutus

Calpurnia

Caesar

1 2 3 5 8 13 21 34

2 4 8 16 32 64 128

13 16

What happens if the word Caesar is added to document 14?

Inverted Index

Linked lists generally preferred to arrays Dynamic space allocation Insertion of terms into documents easy Space overhead of pointers

Brutus

Calpurnia

Caesar

2 4 8 16 32 64 128

2 3 5 8 13 21 34

13 16

1

Dictionary Postings

(sorted by docID)

Inverted Index - Construction

Tokenization

Token Stream Friends Romans Countrymen

Linguistic Preprocessi

ngModified Tokens friend roman countryman

Indexing

Inverted Index

friend

roman

countryman

2 4

2

13 16

1

Documents(to be indexed)

Friends, Romans, countrymen.

Linguistic Preprocessing

Case-folding Often best to lower case everything, since users

will use lowercase regardless of ‘correct’ capitalization.

Removal of stopwords Very common words like the, of, to, etc. List of stopwords (stop lists)

http://www.dcs.gla.ac.uk/idom/ir_resources/ linguistic_utils/stop_words

http://www.ranks.nl/tools/stopwords.html

Linguistic Preprocessing

Stemming Reduce terms to their “roots” before indexing,

e.g., automate(s), automatic, automation all reduced to automat.

Porter’s stemmer Implementations in C, Java, Perl, Python, etc. http://www.tartarus.org/~martin/PorterStemmer/

……

Input a sequence of (term, docID) pairs

I did enact JuliusCaesar I was killed i' the Capitol; Brutus killed me.

Doc 1

So let it be withCaesar. The nobleBrutus hath told youCaesar was ambitious

Doc 2

Term Doc #I 1did 1enact 1julius 1caesar 1I 1was 1killed 1i' 1the 1capitol 1brutus 1killed 1me 1so 2let 2it 2be 2with 2caesar 2the 2noble 2brutus 2hath 2told 2you 2

caesar 2was 2ambitious 2

Indexing

Sort by terms.

Term Doc #ambitious 2be 2brutus 1brutus 2capitol 1caesar 1caesar 2caesar 2did 1enact 1hath 1I 1I 1i' 1it 2julius 1killed 1killed 1let 2me 1noble 2so 2the 1the 2told 2you 2was 1was 2with 2

Term Doc #I 1did 1enact 1julius 1caesar 1I 1was 1killed 1i' 1the 1capitol 1brutus 1killed 1me 1so 2let 2it 2be 2with 2caesar 2the 2noble 2brutus 2hath 2told 2you 2caesar 2was 2ambitious 2

This is the core indexing step,

Merge multiple term entries in a single

document. Add

frequency information.

Term Doc # Freqambitious 2 1be 2 1brutus 1 1brutus 2 1capitol 1 1caesar 1 1caesar 2 2did 1 1enact 1 1hath 2 1I 1 2i' 1 1it 2 1julius 1 1killed 1 2let 2 1me 1 1noble 2 1so 2 1the 1 1the 2 1told 2 1you 2 1was 1 1was 2 1with 2 1

Term Doc #ambitious 2be 2brutus 1brutus 2capitol 1caesar 1caesar 2caesar 2did 1enact 1hath 1I 1I 1i' 1it 2julius 1killed 1killed 1let 2me 1noble 2so 2the 1the 2told 2you 2was 1was 2with 2

Why?Will discuss later.

Split the result into a dictionary file and a postings file.

Doc # Freq2 12 11 12 11 11 12 21 11 12 11 21 12 11 11 22 11 12 12 11 12 12 12 11 12 12 1

Term N docs Tot Freqambitious 1 1be 1 1brutus 2 2capitol 1 1caesar 2 3did 1 1enact 1 1hath 1 1I 1 2i' 1 1it 1 1julius 1 1killed 1 2let 1 1me 1 1noble 1 1so 1 1the 2 2told 1 1you 1 1was 2 2with 1 1

Term Doc # Freqambitious 2 1be 2 1brutus 1 1brutus 2 1capitol 1 1caesar 1 1caesar 2 2did 1 1enact 1 1hath 2 1I 1 2i' 1 1it 2 1julius 1 1killed 1 2let 2 1me 1 1noble 2 1so 2 1the 1 1the 2 1told 2 1you 2 1was 1 1was 2 1with 2 1

Where do we pay in storage?

The index we just built

Doc # Freq2 12 11 12 11 11 12 21 11 12 11 21 12 11 11 22 11 12 12 11 12 12 12 11 12 12 1

Term N docs Tot Freqambitious 1 1be 1 1brutus 2 2capitol 1 1caesar 2 3did 1 1enact 1 1hath 1 1I 1 2i' 1 1it 1 1julius 1 1killed 1 2let 1 1me 1 1noble 1 1so 1 1the 2 2told 1 1you 1 1was 2 2with 1 1

How do we process a query?

Query Processing

Consider processing the query:Brutus AND Caesar Locate Brutus in the dictionary;

Retrieve its postings. Locate Caesar in the dictionary;

Retrieve its postings. “Merge” the two postings:

128

34

2 4 8 16 32 64

1 2 3 5 8 13

21

Brutus

Caesar

34

1282 4 8 16 32 64

1 2 3 5 8 13 21

Query Processing - Merge

128

34

2 4 8 16 32 64

1 2 3 5 8 13 21

Brutus

Caesar2 8

What’s crucial: postings must be sorted by docID.

Walk through the two postings simultaneously, in time linear in the total number of postings entries. In other words, if the posting list lengths are x and y, the merge takes O(x+y) operations.

Query Processing - Exercise

Adapt the merge algorithm for the queries: Brutus AND NOT Caesar Brutus OR NOT Caesar

Can we still run through the merge in time O(x+y)?

Query Processing - Exercise

What about an arbitrary Boolean formula? (Brutus OR Caesar) AND NOT (Antony OR

Cleopatra)

Can we always merge in “linear” time? The time complexity is linear in what?

Can we do better?

Query Processing - Exercise

Extend the merge algorithm 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, then each query term appears only once in the query.

Take Home Messages

IR is about search engines. Search is hot! Search is cool!

Free textbook online! http://www.dcs.bbk.ac.uk/~dell/teaching/ir/

Inverted Index dictionary + postings

Boolean Search query processing