作者: Masashi Toyoda , Masaru Kitsuregawa
关键词:
摘要: Recent advances in storage technology make it possible to store a series of large Web archives. It is now an exciting challenge for us observe evolution the Web. In this paper, we propose method observing web communities. A community set pages created by individuals or associations with common interest on topic. So far, various link analysis techniques have been developed extract We analyze communities comparing four Japanese archives crawled from 1999 2002. Statistics these and are examined, global behavior described. Several metrics introduced measure degree evolution, such as growth rate, novelty, stability. system extracting detailed using metrics. allows understand when how emerged evolved. Some examples shown our system.