Identifying and Ranking Topic Clusters in the Blogosphere
M. Atif Qureshi, Arjumand Younus, Muhammad Subhan Saeed, Nasir Touheed, Emanuele Pianta, Kateryna Tymoshenko · 2010
The blogosphere is a huge collaboratively constructed resource containing diverse and rich information. This diversity and richness presents a significant research challenge to the Information Retrieval community. This paper addresses this challenge by proposing a method for identification of “topic clusters ” within the blogosphere where topic clusters represent the concept of grouping together blogs sharing a common interest i.e. topic, the algorithm takes into account both the hyperlinked social network of blogs along with the content in the blog posts. Additionally we use various forms and parts-of-speech of the topic to provide a broader coverage of the blogosphere. The next step of the method is to assign topic-specific ranks to each blog in the cluster using a metric called “Topic Discussion Rank, ” that helps in identifying the most influential blog for a specific topic. We also perform an experimental evaluation of our method on real blog data and show that the proposed method reaches a high level of accuracy. 1