Observing Evolution of Web Communities

We propose a method for observing evolution of web communities. A web community is a set of web pages created by individuals or associations with a common interest on a topic. So far various techniques have been developed to extract web communities by link analysis. Since a community represents a certain topic, we can understand when and how the topic emerged and evolved in the Web. We have developed a system for observing evolution of web communities by comparing three Japanese web archives periodically crawled in 1999, 2000, and 2001. It rst extracts whole web communities, and their relevances from each archive. Then it provides a web community evolution viewer. The user can observe the evolution, using the relevances and evolution metrics, such as growth rate and novelty. Several evolution examples are shown using our system.