Study of Social Networks
Social networks such as MySpace, Facebook and lately Twitter have become extremely popular with tens or hundreds of millions of people opening accounts and using their accounts to communicate with various purposes and goals in mind.
A current MS student, Beaux Sharifi, has been studying Twitter microblogs and has developed algorithms for summarizing microblogs. An REU student from Fall 2009-Spring 2010, Mark-Anthony Hutton worked closely with Beaux . We were pleased to have two publications out of Beaux and Mark-Anthony's work: NWDAA '10 held at Tezpur, Assam, India and NAACL-HLT '10 to be held in LA in June. We also work with Prof Carmen Stavrositu in Communication whose specialty is study of social media like blogs and microblogs.
There are many ways we want to extend our current work on the study of social networks. Here are some ideas.
- Multi-sentence summarization for Twitter: The current summaries that we produce are single-sentence long. The current approach is to find the central idea in a large number Twitte posts on a trending topic and summarize all of them in a single sentence. However, Twitter post on a topic can have many themes. We want to find the top-ranked themese on a certain topic and obtain a single-sentence summary for each theme on a certain topic.
One way to get started would be to take a bunch of Twitter posts that come back from a search of a topic, cluster them using a document clustering algorithm, and obtain summary for each cluster.
- Visualization of Twitter and other social networking sites’ contents.
What are the relationships among the various trending terms, if any? How do you find the timeline or life cycle of terms and display them?