Communities in Social Media: An Eyepiece into Context, User Intention and Influence

by

Monday, June 30, 2008, 13:00pm - Monday, June 30, 2008, 14:00pm

5-154, JHU APplied Physics Laboratory

community detection, influence, social media, social networking

Communities are central to online social media systems and detecting their structure and membership is critical for many applications. In this talk, I will discuss some of our recent research on both identifying communities and analyzing their content. We leverage the special properties of Social Media data to analyze the communities in an attempt to understand user intentions, context and influence.

Community detection techniques can be computationally expensive. An approach to reducing the cost is by estimating the community structure from only a small fraction of the graph and using spectral methods to identify membership of the rest of the nodes. This technique exploits the fact that in most Web communities a small fraction of the members are highly linked while most (the "long tail") are sparsely connected to the network. It has the advantage of quickly and efficiently finding a reasonable approximation to the community structure of the overall network. I will also briefly describe techniques for identifying communities from feed readership or subscription information.

A recent, new form of communication in Social Media is Microblogging. In microblogging users can describe their current status in short posts distributed by instant messages, mobile phones, email or the Web. We analyze communities in Twitter, a popular microblogging platform and identify associated user intentions. This can be helpful in understanding the rising popularity of microblogging tools.

Blogs, wikis and forums provide a wealth of information for market research. By analyzing the sentiments and opinions expressed about products we can learn about what users like and what are their gripes. I will describe our research on opinion retrieval and its applications in understanding trust and bias in social media communities.

Finally, I will outline some of the practical challenges faced when dealing with social media content. In particular, the problem of spam blogs and effective ways to eliminate them.

OWL Tweet

UMBC ebiquity