Basics of Information Retrieval - Focus: the Web

Lillian N. Cassel

February 2008

For CSC 2500 : Survey of Information Science

A number of these slides are taken or adapted fromSource: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

Basic ideas

Information overload The challenging byproduct of the information age Huge amounts of information available -- how to find

what you need when you need it What kinds of information do you manage?

• What do you expect to find when you want/need it?• How frequently do you access each kind?

Think about addresses, e-mail messages, files of interesting articles, product brochures, business cards, photos, charts and tables, etc.

Information retrieval is the formal study of efficient and effective ways to extract the right bit of information from a collection. The web is a special case, as we will discuss.

Some distinctions

Data, information, knowledge How do you distinguish among them?

Data: symbols Information: data that are processed to be useful; provides

answers to "who", "what", "where", and "when" questions Knowledge: application of data and information; answers

"how" questions Understanding: appreciation of "why” Wisdom: evaluated understanding.

• Russell Ackoff as reported in http://www.systems-thinking.org/dikw/dikw.htm

What are we seeking when we search the web?

Organization

Information sources - organization: Very well organized, indexed, controlled

Give some examples Totally unorganized, uncharacterized,

uncontrolled Give some examples

Something in between Give some examples

Databases Databases hold specific data items

Organization is explicit Keys relate items to each other Queries are constrained, but effective in retrieving the data that is

there Databases generally respond to specific queries with specific

results Browsing is difficult Searching for items not anticipated by the designers can be

difficult Give an example of a database with which you interact regularly.

What is a query that works easily? Have you tried unsuccessfully to get information you know is there?

The Web

The Web contains many kinds of elements Organization? There are no keys to relate items to each other Queries are unconstrained; effectiveness depends on the

tools used. Web queries generally respond to general queries with

specific results Browsing is possible, though somewhat complicated There are no designers of the overall Web structure. Describe how you frequently use the web

What works easily? What has been difficult?

How high is MtEverest

The answer, with source -- not just a link to the place

where an answer could be found

How fast is the Web growing?

Source: http://news.netcraft.com/archives/web_server_survey.html

An increase of 50 million sites during

2007

The servers

What server software is providing content on the Web

Source: http://news.netcraft.com/archives/web_server_survey.html

Digital Library Something in between the very structured

database and the unstructured Web. Content is controlled. Someone makes the

entries. (Maybe a lot of people make the entries, but there are rules for admission.)

Searching and browsing are somewhat open, not controlled by fixed keys and anticipated queries.

Nature of the collection regulates indexing somewhat.

Digital Library examples

American Memory: http://memory.loc.gov/ammem/index.html

The National Science Digital Library www.nsdl.org

CITIDEL: www.citidel.org

In all cases Trying to connect an information user to the

specific information wanted. Concerned with efficiency and effectiveness

Effectiveness - how well did we do? Efficiency - how well did we use available

resources? We will focus on the Web, but some of the

concepts apply in other situations also.

Effectiveness Two measures:

Precision Of the results returned, what percentage are meaningful to the goal of

the query? Recall

Of the materials available that match the query, what percentage were returned?

Ex. Search returns 590,000 responses and 195 are relevant. How well did we do? Not enough information.

Did the 590,000 include all relevant responses? If so, recall is perfect.• We have no way of knowing whether or not this is all the relevant

resources• What is important is whether the right response is there.

195/590,000 is not good precision!

The Web

Web spider

Indexer

Search

User

Web Results 1 - 10 of about 7,310,000 for miele . (0.12 seconds)

Miele, Inc -- Anything else is a compromise At the heart of your home, Appliances by Miele . ... USA. to miele .com. Residential Appliances. Vacuum C leaners. Dishwashers. Cooking Appliances. Steam Oven. Coffee System ... www. miele .com/ - 20k - Cached - Similar pages

Miele Welcome to Miele , the home of the very best appliances and kitchens in the world. www. miele .co.uk/ - 3k - Cached - Similar pages

Miele - Deutscher Hersteller von Einbaugeräten, Hausgeräten ... - [ Translate this page ] Das Portal zum Thema Essen & Geniessen online unter www.zu -tisch.de. Miele weltweit ...ein Leben lang. ... Wählen Sie die Miele Vertretung Ihres Landes. www. miele .de/ - 10k - Cached - Similar pages

Herzlich willkommen bei Miele Österreich - [ Translate this page ] Herzlich willkommen bei Miele Österreich Wenn Sie nicht automatisch weitergeleitet werden, klicken Sie bitte hier! HAUSHALTSGERÄTE ... www. miele .at/ - 3k - Cached - Similar pages

Sponsored Links

CG Appliance Express Discount Appliances (650) 756 -3931 Same Day Certified Installation www.cgappliance.com San Francisco -Oakland-San Jose, CA Miele Vacuum Cleaners Miele Vacuums - Complete Selection Free Shipping! www.vacuums.com Miele Vacuum Cleaners Miele -Free Air shipping! All models. Helpful advice. www.b est -vacuum.com

Source: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

Basic structure of Web Search

What is involved in text retrieval

The process

Query entered

Query Interpreted

Items retrieved

Index searched

Results Ranked

The Collection Where does the collection come from? How is the index created? Those are important distinguishing

characteristics Inverted Index -- Ordered list of terms

related to the collected materials. Each term has an associated pointer to the related material(s). www.cs.cityu.edu.hk/~deng/5286/T51.doc

An example: Unstructured data - circa 1650

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 Ranked retrieval (best documents to return)

Source: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

A table showing incidence of the search terms in the documents

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

1 means the word is in the document

What are the entries in the rows for the query: Brutus AND Caesar NOT Calpurnia

Source: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

The vectors

Brutus: 1 1 0 1 0 0 Caesar: 1 1 0 1 1 1 Calpurnia: 0 1 0 0 0 0 NOT Calpurnia: 1 0 1 1 1 1 Brutus AND Caesar AND NOT Calpurnia: AND them: 1 0 0 1 0 0

Plays # 1 and 4 satisfy the query

The query terms in the documents

Antony 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.

Source: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

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 64128

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 lists

Sorted by docID (more later on why).

Posting

Inverted index construction

Tokenizer

Token stream. Friends Romans Countrymen

Linguistic modules

Modified tokens. friend roman countryman

Indexer

Inverted index.

friend

roman

countryman

2 4

2

13 16

1

More onthese later.

Documents tobe indexed.

Friends, Romans, countrymen.

Sequence of (Modified token, Document ID) pairs.

I did enact JuliusCaesar I was killed

i' the Capitol; Brutus killed me.

Doc 1

So let it be withCaesar. The noble

Brutus 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

Indexer steps

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

Core indexing step.

Multiple term entries in a single document are merged.

Frequency information is added.

Term Doc # Term 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 frequency?Will discuss later.

The result is split 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 Coll 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? Doc # Freq

2 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 Coll 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

Pointers

Terms

Query processing: AND

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

The merge

Walk through the two postings simultaneously, in time linear in the total number of postings entries

128

34

2 4 8 16 32 64

1 2 3 5 8 13 21

Brutus

Caesar2 8

If the list lengths are x and y, the merge takes O(x+y)operations.Crucial: postings sorted by docID.

Boolean queries: Exact match

The 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 terms Views each document as a set of words Is precise: document matches condition or not.

Primary commercial retrieval tool for 3 decades.

Professional searchers (e.g., lawyers) still like Boolean queries: You know exactly what you’re getting.

Example: WestLaw http://www.westlaw.com/

Largest commercial (paying subscribers) legal search service (started 1975; ranking added 1992)

Tens of terabytes of data; 700,000 users Majority of users still use boolean queries Example 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 sentence

Example: WestLaw http://www.westlaw.com/

Another example query: Requirements for disabled people to be able to access

a workplace disabl! /p access! /s work-site work-place

(employment /3 place Note that SPACE is disjunction, not conjunction! Long, precise queries; proximity operators;

incrementally developed; not like web search Professional searchers often like Boolean search:

Precision, transparency and control But that doesn’t mean they actually work better….

Boolean queries: More general merges

Exercise: Adapt the merge for the queries:Brutus AND NOT CaesarBrutus OR NOT Caesar

Query optimization

What is the best order for query processing? Consider a query that is an AND of t terms. For each of the t terms, get its postings, then

AND them together.

Brutus

Calpurnia

Caesar

1 2 3 5 8 16 21 34

2 4 8 16 32 64128

13 16

Query: Brutus AND Calpurnia AND Caesar

Query optimization example

Process in order of increasing freq: start with smallest set, then keep cutting

further.

Brutus

Calpurnia

Caesar

1 2 3 5 8 13 21 34

2 4 8 16 32 64128

13 16

This is why we keptfreq in dictionary

Execute the query as (Caesar AND Brutus) AND Calpurnia.

What’s ahead in IR?Beyond term search

What about phrases? Stanford University

Proximity: Find Gates NEAR Microsoft. Need index to capture position information in

docs. More later. Zones in documents: Find documents with

(author = Ullman) AND (text contains automata).

Evidence accumulation

1 vs. 0 occurrence of a search term 2 vs. 1 occurrence 3 vs. 2 occurrences, etc. Usually more seems better

Need term frequency information in docs

Ranking search results

Boolean queries give inclusion or exclusion of docs.

Often we want to rank/group results Need to measure proximity from query to each

doc. Need to decide whether docs presented to

user are singletons, or a group of docs covering various aspects of the query.

IR vs. databases:Structured vs unstructured data

Structured data tends to refer to information in “tables”Employee Manager Salary

Smith Jones 50000

Chang Smith 60000

50000Ivy Smith

Typically allows numerical range and exact match(for text) queries, e.g.,Salary < 60000 AND Manager = Smith.

Unstructured data

Typically refers to free text Allows

Keyword queries including operators More sophisticated “concept” queries e.g.,

find all web pages dealing with drug abuse

Classic model for searching text documents

Semi-structured data

In fact almost no data is “unstructured” E.g., this slide has distinctly identified zones

such as the Title and Bullets Facilitates “semi-structured” search such as

Title contains data AND Bullets contain search

… to say nothing of linguistic structure

More sophisticated semi-structured search

Title is about Object Oriented Programming AND Author something like stro*rup

where * is the wild-card operator Issues:

how do you process “about”? how do you rank results?

The focus of XML search.

Clustering and classification

Given a set of docs, group them into clusters based on their contents.

Given a set of topics, plus a new doc D, decide which topic(s) D belongs to.

The web and its challenges

Unusual and diverse documents Unusual and diverse users, queries,

information needs Beyond terms, exploit ideas from social

networks link analysis, clickstreams ...

How do search engines work? And how can we make them better?

More sophisticated information retrieval

Cross-language information retrieval Question answering Summarization Text mining …

Crawling the web

Misnomer as the spider or robot does not actually move about the web

Program sends a normal request for the page, just as a browser would. Retrieve the page and parse it.

Look for anchors -- pointers to other pages.• Put them on a list of URLs to visit

Extract key words (possibly all words) to use as index terms related to that page

Take the next URL and do it again Actually, the crawling and processing are parallel

activities

Responding to search queries

Use the query string provided Form a boolean query

Join all words with AND? With OR? Find the related index terms Return the information available about the

pages that correspond to the query terms. Many variations on how to do this. Usually

proprietary to the company.

Making the connections

Stemming Making sure that simple variations in word form are

recognized as equivalent for the purpose of the search: exercise, exercises, exercised, for example.

Indexing A keyword or group of selected words Any word (more general) How to choose the most relevant terms to use as index

elements for a set of documents. Build an inverted file for the chosen index terms.

The Vector model

Let N be the total number of documents in the collection ni be the number of documents which contain ki

freq(i,j) raw frequency of ki within dj

A normalized tf (term frequency) factor is given by tf(i,j) = freq(i,j) / max(freq(i,j)) where the maximum is computed over all terms which occur within

the document dj

The idf (index term frequency) factor is computed as idf(i) = log (N/ni) the log is used to make the values of tf and idf comparable. It

can also be interpreted as the amount of information associated with the term ki.

Anatomy of a web page

Metatags: Information about the page Primary source of indexing information for a search engine. Ex. Title. Never mind what has an H1 tag (though that may

be considered), what is in the <title> </title> brackets? Other tags provide information about the page. This is

easier for the search engine to use than determining the meaning of the text of the page.

Dealing with the cheaters False information provided in the web page to make the

search engine return this page False metatags, invisible words (repeated many times), etc

Standard Metatags

The Dublin Core (http://dublincore.org/)15 common items to use in labeling any web

document

Title Contributor SourceCreator Date LanguageSubject Resources type RelationDescription Format CoveragePublisher Identifier Rights

Hubs and authorities

Hub points to a lot of other places. CITIDEL is a hub for computing information NSDL is a hub for science, technology, engineering and

mathematics education. Authorities are pointed to by a lot of other places.

W3C.org is an authority for information about the web. When Hub or Authority status is captured, the search

can be more accurate. If several pages match a query, and one is an authority

page, it will be ranked higher. When a hub matches a query, the pages it points to are

likely to be relevant.

An exercise We have a document collection (sort of) We build an index so that we can process a query Do this:

Take the list of documents Each group go through the items assigned and list every word

that appears and the number of times the word appears. Use stemming. Only count nouns.

We will then combine these lists and pick the 20 most frequent words to use as the index. Make the inverted list (file index) corresponding to each of the 20 words.

Now, suppose we have the query I will give you. Which documents in our collection satisfy the query? Given our 20 words list, how would we match that to our query?

Conclusions

The plan was to introduce the basic concepts of information retrieval in a form accessible to most students.

There is a lot more, but this gives some flavor of the way the systems work and may help you use them more effectively.

A word about the pattern for these slides …

Extra references

Data, Information, Knowledge, and Wisdom by Gene Bellinger, Durval Castro, Anthony Mills: http://www.systems-thinking.org/dikw/dikw.htm

Some slides from the course CS276 Information Retrieval and Web Mining - Standford University: http://www.stanford.edu/class/cs276/cs276-2006-syllabus.html

http://infolab.stanford.edu/~backrub/google.htmlhttp://www.seochat.com/c/a/Search-Engine-Optimization-Help/Search-Engines-and-Algorithms-Semantic-Search/http://www.columbia.edu/~js1353/pubs/wolf-www02.pdfhttp://mics.cs.uwec.edu/CD%20image/papers/paper89.pdfhttp://www.engl.niu.edu/ehoffman/search_engines.htmlhttp://moon.ouhsc.edu/kboyce/sdms/otherlinks/spidap4.htmhttp://www.cse.lehigh.edu/~brian/course/sem/notes/searchengines.pdfhttp://www9.org/w9cdrom/159/159.htmlhttp://www9.org/w9cdrom/293/293.htmlhttp://perso.fundp.ac.be/~lgoffine/Hypertext/semantic_links.htmlhttp://dbpubs.stanford.edu:8090/pub/2002-6http://searchenginewatch.com/showPage.html?page=2167891http://searchenginewatch.com/showPage.html?page=2168031http://www.webreference.com/content/search/http://www.onlamp.com/pub/a/onlamp/2003/08/21/better_search_engine.html (this site has other related articles links)http://www.webopedia.com/DidYouKnow/Internet/2003/HowWebSearchEnginesWork.asphttp://art-support.com/seo_algorithms.htmhttp://computer.howstuffworks.com/search-engine.htm

Some additional articles on Web search. Note the first one.