Jupiter

1
March 2008
Information Foraging and Scent Following
Samantha Cooper Kristian Gregory Caroline
Ip Susan Lin Rachel Liu
2
Overview of Presentation
Introduction
Information Foraging
Information Scent
Latent Semantic Analysis
Usability
Human Behaviour
Experiment
Conclusion
3
The Changing World of Information
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Introduction
Human Behaviour
Conclusion

• Implications of a growing and developing web
• Unparalleled volume of information
• Unlimited access to information
• Greater importance of acquiring relevant
  information in short time-frames
• (Payne, Howes 2003)

4
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour
Information Retrieval in context The
knowledge acquisition cycle (Payne, Howes 2003)
5
Information Foraging
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour

• History and Roots
• Foraging Theory
• Rational Analysis

6
Foraging Theory
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour

• Organisms forage in such a way to maximise their
  energy intake per unit time
• Patch models How should I allocated my time
  between searching for the best patch and
  searching within a patch?
• Scent models How do I assess the value of food
  from its scent alone?
• Diet models What do I want to eat?
• (Pirolli, Card 1999)

7
Rational Analysis
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour

• Precisely specify goals of cognitive system
• Develop formal model of environment
• Make minimal assumptions about computational
  limitations
• Derive optimal behaviour function given 1-3
• Examine empirical evidence to see if
  predictions of behaviour function are confirmed.
• Refine theory iteratively
• (Chater, Oaksford 1999)

8
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour
How would you go about searching for information
to learn more about your chosen AHCI topic?
9
Information Foraging
Usability
Information Foraging
Latent Semantic Analysis
Introduction
Conclusion
Information Scent
Human Behaviour

• Humans forage in such a way as to maximise the
  relevant information acquired per unit time
• Patch models How do I allocated my time between
  searching for a resource, searching within a
  resource and enrichment
• Scent models How do I assess whether what I
  want lies behind a proximal cue?
• Diet models What resources are useful to me and
  which should I ignore?
• (Pirolli, Card 1999)

10
Information Scent
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Introduction
Conclusion
Human Behaviour
Information scent is the (imperfect) perception
of the value, cost, or access path of information
sources obtained from proximal cues, such as
bibliographic citations, www links, or icons
representing the sources. (Pirolli and Card
1999)

• Information scent is a concept from information
  foraging theory
• If the user feel that they are on the right
  track to the information they require, they will
  continue as long as they sense that they are
  getting warmer i.e. that the scent is getting
  stronger or the user will give up (Nielsen 2003).

11
Information Scent in action
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Conclusion
Introduction
Human Behaviour

• Information Scent is most prevalent when
  browsing a website.
• Approach sites with a specific mission in mind.
  
• First scan for trigger words
• Failure to find appropriate triggers words,
  often resort to Search function
• As users click, scent is more about eliminating
  choices to hone the right content
• If page displayed is inappropriate to users
  needs, user frustration increases and will
  consider leaving.

12
Weak Information Scent
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Conclusion
Introduction
Human Behaviour
13
ButStrong information scent can backfire
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Conclusion
Introduction
Human Behaviour
14
Semantically Similar Words
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Conclusion
Introduction
Human Behaviour
What words can we use to describe this picture?
15
The Resulting Problem..
Usability
Information Foraging
Information Scent
Latent Semantic Analysis
Conclusion
Introduction
Human Behaviour

• Different words with the same meaning
• A word with multiple meanings
• An individuals interpretation

• Research has found that the same word is used by
  two people to describe an object
• 10-20 of the time
• (Peter Foltz, 1990)

• Significant impact on effective information
  retrieval

16
Latent Semantic Analysis
Usability
Information Foraging
Latent Semantic Analysis
Conclusion
Introduction
Information Scent
Human Behaviour

• Developed by Bell Communication Research (1988)
  to address early information retrieval systems
  that performed exact word matching
• It computes the semantic similarity between
  words

• One particular example of where LSA is used is
  accessing usability problems on a website

17
Use of words for navigational design
Usability
Information Foraging
Latent Semantic Analysis
Conclusion
Introduction
Information Scent
Human Behaviour

• Ensure
• Links
• Category descriptions
• Menu items
• Breadcrumbs

Explicitly describe what the user will find at
the destination
18
Human Behaviour
Latent Semantic Analysis
Usability
Conclusion
Information Foraging
Introduction
Information Scent
What is Usability?

• The effectiveness, efficiency and satisfaction
  with which users can achieve tasks in particular
  environment of a product

19
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent
What is Usability?

• Web usability main components
• Learnable
• Effective
• Memorable
• Reliable
• Enjoyable

20
Role of Usability in the Web
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Stickiness of website determines popularity and
  usefulness
• Positive correlation between usability and
  information scent
• Good navigation scent retrieves content quickly

21
Design Guidelines for Usability
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Keep text as short and concise
• Keep links accurate and unambiguous
• Use simple headings, page titles and page
  structure
• Back buttons and home links accessible
• Listen to users and allow users to send feedback
  easily

22
Worse Design Mistakes
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent
a) Anything that looks like an Advertisement
b) Cluttered text
c) Violating design conventions
d) Not changing the colour of visited links
23
What are Breadcrumbs?
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• A method to provide good navigation scent
• Metaphor of Hansel and Gretel
• 3 Types
• Path
• Attribute
• Location
• Home page -gt Section Page -gt Subsection page

24
What of breadcrumbs used have a horizontal
orientation?
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent
25
Role of Breadcrumbs
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Breadcrumbs should show the site hierarchy not
  the users history
• Hypothetical benefits of breadcrumbs
• Improves site efficiency
• Good signpost especially in large websites
• Perform one-click access to higher-level site

26
Experiment
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Usage effectiveness of breadcrumbs
• Do users choose to use this navigation method?
• If so, are they really helpful?

27
Why is human behaviour important?
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Majority of users do not type in more than one
  search query during a session on a website
• Peoples information needs form the foundation
  of their goals in searching for information and
  the methods they use to find information on the
  web.
• Greater understanding of user behaviour can
  lead to better support in tasks users perform in
  web environments Kellar06.

28
What drives users information needs and
behaviours?
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Previous studies have shown that users develop
  patterns in the way they search for information
• Card et al investigations around user
  navigation behaviour suggested that there are
  individual differences in the way humans complete
  the same tasks.
• These individual differences include
• Individual interests
• Users ability
• Experience
• Cognitive and psychological differences
• Holists and Serialists

29
Holist Learners Serialist Learners Gordon
Pask
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent
30
Users Information Seeking Behaviours
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent

• Many researchers have conducted studies with
  the goal to identify and define users
  information seeking behaviours on the web
  Choo99.
• Pirolli et al explored the methods used for
  information foraging, and categorised users
  browsing methods into four main types
• Exploring
• Monitoring
• Finding
• Collecting

31
Percentage of time spent on activities
Human Behaviour
Conclusion
Latent Semantic Analysis
Usability
Information Foraging
Introduction
Information Scent
32
Conclusion
Human Behaviour
Information Foraging
Scent Following
Latent Semantic Analysis
Usability
Conclusion
Human Behaviour

• Studies with larger sample sizes for more
  generalisable empirical results
• Attempts to generate foraging models based on
  weaker assumptions
• Customisable search toolkits

33
Questions
34
References
CATLEDGE, L. J.E, P. (1997), Characterizing
browsing strategies in the world-wide web, /in
/'Computer Networks and ISDN Systems' CHATER,
N., OAKSFORD, M., 1999. Ten Years of the Rational
Analysis ofCognition. Trends in Cognitive
Sciences. Vol 3 (2), pp.57-65. CROFT.B.
Improving the effectiveness of information
retrieval with local context analysis. ACM
transactions on information systems, Jan
2000. NIELSEN, J. Top 10 Mistakes online
(http//www.useit.com/alertbox/9605.html) NIELSEN
. J. Ten Good Deeds in Web Design online
(http//www.useit.com/alertbox/991003.html) NIELS
EN. J. Top Ten Mistakes in Web Design online
(http//www.useit.com/alertbox/9605.html) OATES.T
BHAT.V Using Latent Semantic Analysis to find
different names for the same entity in free text.
Proceedings of the 4th international workshop on
Web information and data management, Nov 2002.
PAYNE, S.,HOWES, A., DIX, A., 2003. Post-web
cognition evolving knowledgestrategies for
global information environments. International
Journal of WebEngineering and Technology 1 (1)
pp.112-126. PIROLLI, P., CARD, S.K., 1999.
Information foraging. Psychological
Review,106(4), pp.643-675.