Fawaz Ghali

1
Web 2.0 for the Adaptive Web

• Fawaz Ghali
• http//www.fawazghali.com/

2
Overview

• Web 2.0
• User Profile in Web2.0
• Recommending Systems
• Content-based Filtering
• Collaborative Filtering
• Hybrid Filtering
• Recommendations to Groups
• Social Filtering

3
Web 2.0
4
User Profile in Web 2.0

• Examples of explicit data collection
• Asking a user to rate an item on a sliding scale.
• Asking a user to rank a collection of items from
  favourite to least favourite.
• Presenting two items to a user and asking him/her
  to choose the best one.
• Asking a user to create a list of items that
  he/she likes.

5
User Profile in Web 2.0

• Examples of implicit data collection
• Observing the items that a user views in an
  online store.
• Analyzing item/user viewing times
• Keeping a record of the items that a user
  purchases online.
• Obtaining a list of items that a user has
  listened to or watched on his/her computer.
• Analyzing the user's social network and
  discovering similar likes and dislikes

6
Recommender Systems

• Specific type of information filtering (IF)
  technique that attempts to present information
  items (movies, music, books) that are likely of
  interest to the user.
• Comparing the user's profile to some reference
  characteristics.
• These characteristics may be from the information
  item (content-based approach) or the user's
  social environment (collaborative filtering).

7
Content-based Filtering

• Items are used as parameters instead of users.
• Grouping various items together in groups so
  consumers can compare them all together.
• Users use and test the item and give it a rating
  that is relevant to the item and the item class.

8
Content-based Filtering

• The items are classified based on the rating.
• The items are used and tested by the same user or
  group in order to get an accurate rating.
• More reading http//www.dcs.warwick.ac.uk/acrist
  ea/courses/CS411/2008/Book20-20The20Adaptive20
  Web/Content-basedRecommendationSystems.pdf

9
Collaborative Filtering

• Collaborative filtering is the process of
  filtering information or patterns using
  techniques involving collaboration among multiple
  users.
• The method of making automatic filtering about
  the interests of a user by collecting taste
  information from many users (collaborating).

10
How It Works?

• Collaborative filtering systems usually take two
  steps
• 1. Look for users who share the same rating
  patterns with the active user (the user whom the
  prediction is for).
• 2. Use the ratings from those like-minded users
  found in step 1 to calculate a prediction for the
  active user.

11
Active Collaborative Filtering

• Peer-to-peer approach peers, co-workers, and
  people with similar interests rate products,
  reports, and other material objects, also sharing
  this information over the web for other people to
  see.

12
Active Collaborative Filtering

• Advantages Actual rating helps to determine the
  value of the item, find related items on hand.
• Disadvantage The opinion may be biased,
  requires action by the user, user expectations
  may not be met.

13
Passive Collaborative Filtering

• Collects information implicitly.
• A web browser is used to record a users
  preferences by following and measuring their
  actions.
• These implicit filters are then used to
  determine what else the user will like and
  recommend potential items of interest.

14
Explicit vs. Implicit filtering

• Within active and passive filtering there are
  explicit and implicit methods for determining
  user preferences.
• Explicit collection of user preferences requires
  the evaluator to indicate a value for the content
  on a rating scale.
• Implicit collection does not involve the direct
  input of opinion by the user, but instead it is
  assumed that their opinion is implied by their
  actions (reduces the demand on the user, which
  can mean that much more data is available)

15
Collaborative Filtering Problems

• The First-Rater Problem is caused by new items.
  The system is unable to generate semantic
  interconnections to these items and therefore are
  never recommended.
• The Cold-Start Problem is caused by new users in
  the system which have not submitted any ratings.
  Without any information about the user the system
  is not able to guess the user's preferences and
  generate recommendations.

16
Hybrid Recommending System

• A combination of multiple recommending
  techniques.
• Example Collaborative filtering and
  content-based techniques.
• More reading
• http//www.dcs.warwick.ac.uk/acristea/courses/CS
  411/2008/Book20-20The20Adaptive20Web/HybridWeb
  RecommenderSystems.pdf

17
Recommendations to Groups

• Often the users work in groups.
• Web 2.0 phenomena.
• 1. acquiring information about the users
  preferences
• 2. generating recommendations
• 3. explaining recommendations
• 4. helping users to settle on a final result.

18
Recommendations to Groups

• Acquiring information about the users
  preferences.
• If users specify their preferences explicitly, it
  may be desirable for them to be able to examine
  each others preference.
• What benefits and drawbacks can such examination
  have, and how can it be supported by the system?

19
Recommendations to Groups

• The system generates recommendations.
• Some procedure for predicting the suitability of
  items for a group as a whole must be applied.
• What conditions might such a procedure be
  required to fulfil?

20
Recommendations to Groups

• The system presents recommendations to the
  members.
• The suitability of a solution for the individual
  members becomes an important aspect of a
  solution.
• How can relevant information about suitability
  for individual members be presented effectively?

21
Recommendations to Groups

• The system helps the members to reach to
  agreement on which recommendation (if any) to
  accept.
• The final decision is not necessarily made by a
  single person negotiation may be required. How
  can the system facilitate the necessary
  communication among group members?
• More reading http//www.dcs.warwick.ac.uk/acrist
  ea/courses/CS411/2008/Book20-20The20Adaptive20
  Web/RecommendationGroups.pdf

22
Social Filtering Systems

• Bring users together to satisfy explicit
  information needs, or interpersonal interests.
• Compute the similarity between users or groups,
  given their interests or information needs.
• More info http//www.dcs.warwick.ac.uk/acristea/
  courses/CS411/2008/Book20-20The20Adaptive20Web
  /AdaptiveSupportDistributedCollaboration.pdf

23
Aggregation of Ratings for Individuals

• For each candidate item ci and each member mj ,
  the system can predict how mj would evaluate (or
  rate) cj if he or she were familiar with it
• 1. For each candidate ci
• For each member mj predict the rating rij of
  ci
• Compute an aggregate rating Ri from the set
  rij.
• 2. Recommend the set of candidates with the
  highest predicted ratings Ri.

24
Example POLYLENS
25
Review

• Web 2.0
• User Profile in Web2.0
• Recommending Systems
• Content-based Filtering
• Collaborative Filtering
• Hybrid Filtering
• Recommendations to Groups
• Social Filtering

26
Lab session Thursday 12 Nov

• Demonstration of Web 2.0 techniques for the
  Adaptive Web
• Collaborative authoring authoring for
  collaboration group-based adaptive authoring
  social annotation recommender authoring tool MOT
  2.0
• Bring your laptop!

27
Questions and Ideas

• F.Ghali_at_warwick.ac.uk
• DCS, Room 318