Skip to Main Content
The rapid development of blogs has brought on some serious problems such as disclosure of sensitive information, spread of unhealthy information, etc. So it is very important for supervisors to detect them. The common methods based on search engines have some drawbacks such as lower efficiency and lower precision because they need to retrieve and update blog pages frequently, and to analyze all blog pages downloaded In this paper, we present a new method for detecting contents of blogs based on communities discovery. It adopts a hierarchical clustering algorithm based on multi-attribute metrics. By this method, we can monitor the whole blog community by detecting the contents of eigenvalue nodes . Our experimental result shows that the algorithm proposed is highly effective.
Date of Conference: 12-14 Oct. 2008