Important event extraction of chinese document based on small world model

A document is represented as an event network, the nodes represent events, and the edges represent the correlations between events. Our experiment shows that the event network has the small-world properties, i.e., highly clustered and short path length. We pick out important events according to event's importance to make the event network small. Our experiment shows that these important events can make a good summarization of the document. Although there is much research on text processing and applications based on small-world, it is novel to analyze document based on event with the small-world properties of event network.