Social Systems: Can We Do More Than Just Poke Friends

Download Report

Transcript Social Systems: Can We Do More Than Just Poke Friends

XUE Tianfan
The Chinese University of Hong Kong
This work is by Georgia Koutrika, published on CIDR'09
All the figures & tables in these slides are from that paper
Outline
 Motivation
 CourseRank
 Unique features
 Lessons Learnt so Far
 Interaction with rich data
 Conclusion
Motivation – CourseRank
Motivation
 Social Web Site
 FaceBook, del.icio.us, Y! Answer, Flickr, MySpace
 Great success
 Is it interesting for research community?
 Are there any interesting chanllenges to researchers?
 Can we do more than just poke friends?
Motivation
 Social Web Site V.S. Traditional Open Web V.S. Database
 Social Web Site
- mostly unstructured
- Centrally stored
- Users-to-Users Access Control
 Traditional Open Web
- Unstructured
- highly distributed in storage
- Many provider and consumers without access control
 Database
- Structured
- Centrally stored
- 1 provider, many consumers
Motivation
 Social Web Site V.S. Traditional Open Web V.S. Database
Motivation
 Research topics in database
 Research topics in Web search
 What is important for social website
 What is most effective way for users to interact?
 What can be shared among the users?
 What information can be trusted?
 How users to visualize and interact with information?
 How users interact with other users?
 How system evolve over time?
CourseRank
 CourseRank
 An educational social site where Stanford students can
explore course offerings and plan their academic program
 Describe the insight of CourseRank in this paper
CourseRank
 What CourseRank can do
 Search for courses
 Rank courses
 Requirement check
 Feedback to faculties
 etc.
CourseRank
 Unique features
 Hybrid system – database + social system
 Rich data
 New tools – plannar, requirement checker, CourseCloud, etc.
 Site Control
 Closed Community & Restricted Access
 Constituents
Lessons Learnt so Far
 Lessons Learnt so Far
 Meaningful Incentives
-
-
Yahoo! Answers:
Best answer – 10 points, vote for best answer – 1 point
CourseRank:
Different tools: planner, Q&A forum seeds
 Interaction for Constituents
- Department Requirement
both useful for staff and students
Lessons Learnt so Far
 Lessons Learnt so Far
 Meaningful Incentives
-
-
Yahoo! Answers:
Best answer – 10 points, vote for best answer – 1 point
CourseRank:
Different tools: planner, Q&A forum seeds
 Interaction for Constituents
- Department Requirement
both useful for staff and students
Lessons Learnt so Far
 Lessons Learnt so Far
 The power of a closed community
-
-
Block spammers and malicious users
User are more willing to contribute
Example: group forum, department forum, school forum, public
forum
 It’s the Data, Stupid
-
-
External data
Hard to be shared data
Lessons Learnt so Far
 Lessons Learnt so Far
 Privacy can be “shared”
-
The course planned to be taken of a student -> closed
community
 Closed Loop Feedback
-
Build by stanford students theirself, quickly get feedback
 Beyond CourseRank: The Corporate Social Site
-
Example: Inner forum of a company
Can corporate social site learn something from CourseRank?
Interaction with Rich Data
 Rich data
 A student want to take a course:
Course name&description, user’s profile(major, class,
grade), course interrelationships, user’s comments, etc.
 Problem of typical search engines
 a student want something related to Greece


Search “Greece” -> no result
Search “Greek, science” -> got the course “history of science”
 Search engine does not provide user specific result

“Java” is a good course, but not fit for non-engineering students
Interaction with Rich Data
 Data Clouds
 A data cloud is a tag cloud, where the “tags” are the most
representative or significant words found in the results of a
keyword search over the database.
 Example:
“American” -> “Latin American”, “Indians”, and “politics”.
“American”: 1160 courses
“Latin American”: 123 courses
 Challenge:



Multiple relation: tags does not only appear in course name and
description. For example, “java”.
How to rank the result
How to dynamically and efficiently update cloud
Interaction with Rich Data
 Data Clouds
Interaction with Rich Data
 Flexible Recommendation (FlexRecs)
 Why
 Provide recommendation is not easy considering multiple
connections. It need to be manually adjusted.
 Previous recommendation algorithm is fixed
Interaction with Rich Data
 Flexible Recommendation Example
 Relations:
 Simple reconmmendation example
Interaction with Rich Data
 Flexible Recommendation Example
 Complicated reconmmendation example
: recommend
: Expand
: Select
: Connect
Conclusion
 Social sites:
 A closed, well defined community
 Provide rich data
 Not simply for sharing links and networkings
 Two mining tools
 Data clouds
 FlexRecs
Q&A