Ranking Weblogs by Analyzing Reading and Commenting Activities

In this paper, we analyze people’s reading and commenting behaviors in blogspace and proposed an algorithm for blog ranking. Upon two selected communities, AI and Medical, we show how comments, reading records, active browsing and multi time browsing can help to construct the weblog graph and reflect a blog’s popularity. Based on these analysis, we propose cRank, a graph based algorithm, to rank blog among community members. Finally, we divide our dataset temporally and present how the proposed algorithm can make prediction on blogs’ rankings. The experiment shows that cRank has a better performance upon several baseline systems.