Lecture 3

Retrieval Models
The Information Retrieval Cycle

Source
Selection Resource
Query
Formulation Query
Search Ranked List
Selection Documents
System discovery
Vocabulary discovery
Concept discovery
Document discovery Examination Documents
source reselection
Delivery
What is a model?
 A model is a construct designed help us
understand a complex system
 A particular way of “looking at things”
 Models inevitably make simplifying assumptions
 What are the limitations of the model?
 Different types of models:
 Conceptual models
 Physical analog models
 Mathematical models
 …
The Central Problem in IR
Information Seeker Authors
Concepts Concepts
Query Terms Document Terms
Do these represent the same concepts?

The IR Black Box
Query Documents
Representation Representation
Function Function
Query Representation Document Representation
Comparison
Function Index
Hits
Models
 Boolean model
 Based on the notion of sets
 Documents are retrieved only if they satisfy Boolean
conditions specified in the query
 Does not impose a ranking on retrieved documents
 Exact match
 Vector space model
 Based on geometry, the notion of vectors in high
dimensional space
 Documents are ranked based on their similarity to the
query (ranked retrieval)
 Best/partial match
 Probabilistic Model (Language model)
 Based on the notion of probabilities and processes for
generating text
 Documents are ranked based on the probability that
they generated the query
 Best/partial match
Representing Text
Query Documents
Representation Representation
Function Function
Query Representation Document Representation
Comparison
Function Index
Hits
How do we represent text?
 How do we represent the complexities of
language?
 Keeping in mind that computers don’t “understand”
documents or queries
 Simple, yet effective approach: “bag of words”
 Treat all the words in a document as index terms for
that document
 Assign a “weight” to each term based on its
“importance”
 Disregard order, structure, meaning, etc. of the words
What’s a “word”? We’ll return to this in a few lectures…

Sample Document
McDonald's slims down spuds 16 × said
Fast-food chain to reduce certain types of
fat in its french fries with new cooking oil. 14 × McDonalds
NEW YORK (CNN/Money) - McDonald's Corp. is
cutting the amount of "bad" fat in its french fries 12 × fat
nearly in half, the fast-food chain said Tuesday as
it moves to make all its fried menu items
healthier. 11 × fries
But does that mean the popular shoestring fries
won't taste the same? The company says no. "It's 8 × new
a win-win for our customers because they are
getting the same great french-fry taste along with
an even healthier nutrition profile," said Mike 6 × company french nutrition
Roberts, president of McDonald's USA.
But others are not so sure. McDonald's will not 5 × food oil percent reduce
specifically discuss the kind of oil it plans to use,
but at least one nutrition expert says playing with taste Tuesday
the formula could mean a different taste.
Shares of Oak Brook, Ill.-based McDonald's …
(MCD: down $0.54 to $23.22, Research,
Estimates) were lower Tuesday afternoon. It was
unclear Tuesday whether competitors Burger
King and Wendy's International (WEN: down
$0.80 to $34.91, Research, Estimates) would
follow suit. Neither company could immediately “Bag of Words”
be reached for comment.
…
What’s the point?
 Retrieving relevant information is hard!
 Evolving, ambiguous user needs, context, etc.
 Complexities of language
 To operationalize information retrieval, we must
vastly simplify the picture
 Bag-of-words approach:
 Information retrieval is all (and only) about matching
words in documents with words in queries
 Obviously, not true…
 But it works pretty well!
Why does “bag of words” work?
 Words alone tell us a lot about content
Random: beating takes points falling another Delhi 355

Alphabetical: 355 another beating Delhi falling points
“Interesting”: Delhi points beating falling 355 another
Actual: Delhi takes another beating, falling 355 points
 It is relatively easy to come up with words that

describe an information need
Boolean Retrieval
 Weights assigned to terms are either “0” or “1”
 “0” represents “absence”: term isn’t in the document
 “1” represents “presence”: term is in the document
 Build queries by combining terms with Boolean
operators
 AND, OR, NOT
 The system returns all documents that satisfy the
query
Why do we say that Boolean retrieval is “set-based”?

AND/OR/NOT
All documents
A B
C
Logic Tables
B 0 1
A B 0 1
0 0 1 1 0
1 1 1 NOT B
A OR B
B B
A 0 1 A 0 1
0 0 0 0 0 0
1 0 1 1 1 0
A AND B A NOT B
(= A AND NOT B)
Boolean Retrieval
 Advantages
 Results are predictable, relatively easy to explain
 Many different features can be incorporated
 Efficient processing since many documents can be
eliminated from search
 Disadvantages
 Effectiveness depends entirely on user
 Simple queries usually don’t work well
 Complex queries are difficult
Vector Representation
 “Bags of words” can be represented as vectors
 Why? Computational efficiency, ease of manipulation
 Geometric metaphor: “arrows”
 A vector is a set of values recorded in any
consistent order
“The quick brown fox jumped over the lazy dog’s back”
[111111112]
1st position corresponds to “back”
2nd position corresponds to “brown”
3rd position corresponds to “dog”
4th position corresponds to “fox”
5th position corresponds to “jump”
6th position corresponds to “lazy”
7th position corresponds to “over”
8th position corresponds to “quick”
9th position corresponds to “the”
Representing Documents
Document 1
Document 2
Document 1
The quick brown Term

fox jumped over
the lazy dog’s aid 0 1
back. all 0 1 Stopword
back 1 0
brown 1 0
List
come 0 1 for
dog 1 0 is
fox 1 0 of
good 0 1
Document 2 the
jump 1 0 to
lazy 1 0
Now is the time men 0 1
for all good men now 0 1
to come to the over 1 0
aid of their party.
party 0 1
quick 1 0
their 0 1
time 0 1
 3-d pictures useful, but can be misleading for
high-dimensional space
Vector Space Model
 Advantages
 Simple computational framework for ranking
 Any similarity measure or term weighting scheme could
be used
 Disadvantages
 Assumption of term independence
 No predictions about techniques for effective ranking
Probabilistic Model (Language
model)
 Robertson (1977)
 “If a reference retrieval system’s response to each
request is a ranking of the documents in the collection
in order of decreasing probability of relevance to the
user who submitted the request,
 where the probabilities are estimated as accurately as
possible on the basis of whatever data have been
made available to the system for this purpose,
 the overall effectiveness of the system to its user will
be the best that is obtainable on the basis of those
data.”
IR as Classification

Lecture 3

Uploaded by

Copyright:

Available Formats

Lecture 3

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 3

Uploaded by

Copyright:

Available Formats

Retrieval Models

The Information Retrieval Cycle

Search Ranked List

Query Terms Document Terms

Do these represent the same concepts?

Query Representation Document Representation

Query Representation Document Representation

What’s a “word”? We’ll return to this in a few lectures…

Random: beating takes points falling another Delhi 355

 It is relatively easy to come up with words that

Why do we say that Boolean retrieval is “set-based”?

The quick brown Term

You might also like