2. Agenda
• Motivation
• Problem statement
• Methodology
• Concept
• Implementation
• Evaluation
• Conclusion and future work
3. Motivation
• Web 2.0
• User generated content
• Social Networks
• Microblogging
• Twitter
http://blog.socialmaximizer.com/wp-content/uploads/2012/09/Social-Media.jpg
4. Motivation
• 57% of people talk to people more online than they do in real life
• 40% of Twitter users don’t tweet, but instead use it to keep up to date
• A great majority of tweets are just 40 characters long
• Social media use is becoming much more even across age groups (see graph below)
http://thesocialskinny.com/100-social-media-statistics-for-2012/
5. Motivation ctd.
• Huge amount of informations
• Sharing of interests, experiences etc.
• no cultural or georgraphical boundaries
• Implicit knowledge
• Appliances: conferences, course support, viral
marketing
6. Problem statement
• Cluster users into sub-networks based upon their interest
using topic items and social relations
• Provide a filtered view on information generated in their
micro sub-networks
• Which methods or technologies would be suitable for this
challenge?
• Define and evaluate the metrics that can be used to achieve
this goal!
16. Conclusion and future work
• Results encouraging but:
– More accurate and qualitative evaluation of
clustering
– Involving other methods Pearson, Jaccard
– Extending the measurement on more appliance
cases and reference users regarding the
collaborative learning issues
• Later: k-means, hierarchical clustering