Nitin Agarwal, Huan Liu, John J. Salerno, Sanjay Sundarajan
Social interactions are an essential ingredient of our lives. People convene groups and share views, opinions, thoughts, and perspectives. Similar tendencies for social behavior are observed in the World Wide Web. This inspires us to study and understand social interactions evolving in online communities especially in the blogosphere. In this paper, we study and analyze various interaction patterns in community blogs. This would lead to better understanding of the socio-cultural ties between these communities to foster collaboration, better personalization, predictive modeling, and enable tracking and monitoring. Tapping community interactions via link analysis has its limitations due to exponentially large search space. We propose a model, circumventing the challenges with link analysis based approach, to observe interaction within community blogs via an observed event and community reaction to that by studying the opinion and sentiments of the members towards that event. We present a case study on ethnic community blogs exploiting the proposed model and report our findings and observations. During our study we encountered several challenges with the proposed model. We discuss these issues and present future directions to make the model more robust.
Subjects: 1.10 Information Retrieval; 10. Knowledge Acquisition
Submitted: Jul 3, 2008