IDENTIFYING THE LARGE-SCALE STRUCTURE OF THE BLOGOSPHERE
Makoto Uchida (),
Naoki Shibata (),
Yuya Kajikawa (),
Yoshiyuki Takeda (),
Susumu Shirayama () and
Katsumori Matsushima ()
Additional contact information Makoto Uchida: School of Engineering, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8568, Japan
Naoki Shibata: School of Engineering, The University of Tokyo, 2-11-16 Yayoi, Bunkyo, Tokyo 113-8656, Japan
Yuya Kajikawa: School of Engineering, The University of Tokyo, 2-11-16 Yayoi, Bunkyo, Tokyo 113-8656, Japan
Yoshiyuki Takeda: School of Engineering, The University of Tokyo, 2-11-16 Yayoi, Bunkyo, Tokyo 113-8656, Japan
Susumu Shirayama: Research into Artifacts, Center for Engineering (RACE), The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8568, Japan
Katsumori Matsushima: School of Engineering, The University of Tokyo, 2-11-16 Yayoi, Bunkyo, Tokyo 113-8656, Japan
Abstract:
We analyze a topological structure of networks formed according to the entries and trackbacks in the blogosphere, which is a collection of weblog articles. The analysis is performed based on community extraction, network visualization and keyword analysis. It is shown that the large-scale structure of the blogosphere has a globally sparse, but locally dense structure. The entries in a community yield a dense structure while the trackbacks that interconnect communities are sparse. The visualized results show sparkling-firework-like patterns. We then attempt to characterize the communities using a tf-idf technique. It is found that specific topics are discussed in each community. These results will help us to identify the communities in which certain specific topics discussed and to detect trends in the blogosphere.