Information%20seeking

1
Information seeking

• Information-seeking models
• Search strategies
• Search tactics

2
Information-seeking (IS) models

• Holistic view of IS process
• Possible types of IS model
• User-centred
• Descriptive
• Process-oriented
• System-centred
• Prescriptive
• Object-oriented
• Other

3
User-centred IS models

• Description of IS stages / processes
• May aid construction of supportive interface
• Examples
• Simple IS model
• Kuhlthaus ISP model

4
Simple IS model

• Formulate information need
• Identify information sources / channels
• Search for information
• Formulate and submit query
• Examine and evaluate results of search
• Iteration of search stages if necessary

5
Limitations of simple IS model

• Not appropriate for browsing systems
• Assumption of static information need
• Overwhelming emphasis on search itself

6
ISP model 1

• Intended as a search strategy teaching tool for
  expert intermediaries
• Modelling of variations in user uncertainty
• Dimensions
• Affective (emotional)
• Cognitive (thoughts)
• Physical (actions)

7
ISP model 2

• Problem formulation
• Initiation
• Awareness of lack of knowledge and understanding
• Attempt to understand task and relate it to prior
  knowledge
• Selection
• Identification and selection of topic to be
  investigated

8
ISP model 3

• Exploration
• Investigation of information on general topic
• Formulation
• Formation of focussed perspective on the topic
• Development of ability to identify relevant
  information

9
ISP model 4

• Problem solving
• Collection
• Specification of need for relevant, focussed
  information
• Presentation
• Completion of search and usage of results

10
Limitations of ISP model

• Not intended for browsing
• Model has been partially validated for hypertext
  environment by another researcher
• Highly idealised
• Sequential (no iteration)

11
System-centred IS models

• Description of desirable system functions
• Aid in construction of intelligent systems
• Example
• Belkins MONSTRAT model

12
MONSTRAT model 1

• Based on cognitive model of IR interaction
• Models
• System characteristics
• User characteristics
• Problem characteristics
• Ten functions which correspond to system modules

13
MONSTRAT model 2

• Dialogue mode
• Determine appropriate dialogue type for situation
  
• Problem state
• Determine role of user in problem treatment
  process
• Problem mode
• Determine appropriate system capability
• User model
• Generate description of user type, goals, beliefs

14
MONSTRAT model 3

• Problem description
• Generate description of problem type, topic,
  structure, environment, etc
• Retrieval strategy
• Choose and apply appropriate retrieval strategies
  to knowledge resource
• Response generator
• Determine structure of response appropriate to
  user and situation

15
MONSTRAT model 4

• Input analyst
• Convert input from user into structures usable by
  functional experts
• Output generator
• Convert response to form appropriate to user and
  situation
• Explanation
• Describe system operation, capabilities etc to
  user

16
Other IS models 1

• Bates berry-picking model
• Suitable for browsing
• Information needs are dynamic
• Knowledge is gathered throughout process
• Implications for interface design

17
Other IS models 2

• Reids task-oriented model
• Model the broader IS context
• Work task
• Contextual factors
• Social factors

18
Other IS models 3
19
Search strategies

• General strategies
• Overall plan for the search session
• Different strategies for different access types
• Query-based
• Hypermedia
• Term selection strategies
• Specific strategies
• For individual collections, systems, thesauri, etc

20
Query-based strategies 1

• Starting strategies
• Select
• Break complex query into topics and deal with
  each topic separately
• Exhaust
• Include most elements of the query in the initial
  query formulation

21
Query-based strategies 2

• Continuation strategies
• Building blocks
• Combination of discrete topics
• Pearl growing
• Small relevant set expanded gradually
• Successive fractions
• Large relevant set refined gradually

22
Hypermedia strategies

• Strategies often used in combination
• Scanning (information structure)
• Browsing (casual, undirected exploration)
• Selection (choice of individual elements)
• Navigation (chain of scan and select operations)

23
Term selection strategies

• Strategies employed by expert searchers depend
  on
• Vocabulary - free-text vs controlled
• Current state of search process
• Number of documents retrieved
• NLP functionality, e.g. use of proper names
• Used as the basis of expert system rules for
  query reformulation

24
Search tactics

• Individual actions taken at a search stage
• Three possible steps
• Term tactics
• Choose a source of new terms, e.g. thesaurus,
  WordNet, terms from relevant documents
• Search formulation tactics
• Design or redesign the search formulation
• Idea tactics
• Provide ideas to change search direction

25
Search formulation tactics 1

• Exhaust
• Add components to the query
• Reduce
• Remove components from the query
• Union
• Specify union of 2 sets representing different
  query components

26
Search formulation tactics 2

• Intersect
• Specify intersection of 2 sets representing
  different query components
• Parallel
• Include synonyms or conceptually similar terms
• Vary
• Alter / substitute some of the search terms

27
Search formulation tactics 3

• Block / negate
• Reject items containing, or indexed by, certain
  terms
• Neighbour
• Add additional neighbouring terms from current
  document

28
Search formulation tactics 4

• Trace
• Examine documents already retrieved for new terms
  
• Fix
• Try alternative affixes

29
Idea tactics 1

• Skip
• Shift view of the query laterally
• Shift focus from one part of a complex query to
  another
• View the query from a different conceptual angle
• Focus
• Take a narrower perspective of the query
• Choose a limited subset of the query terms
• Fix on a limited conceptualisation of the query

30
Idea tactics 2

• Limit
• Limit the search
• Specify constraints, e.g. for language, data set,
  publication year, etc

31
Summary

• Information-seeking models
• System-centred
• User-centred
• Other, e.g. task-oriented
• Search strategies
• Overall plan for the search session
• Search tactics
• Individual actions taken at a search stage