Discovering Evolution of Complex Event Based on Correlations Between Events

Abstract

There are large numbers of news articles on Web pages every day. Each article usually reports some aspects of a complex event, but they do not report the whole picture of the event. People are often interested in not only a single event but also the correlations between events and the evolution of the complex event, if they want to be aware of the whole picture of the complex event, they have to browse many Web pages. To solve this problem, we propose a method to find the correlations between events, and the evolution of the complex events. We use the signal words and the co-occurrence of the events in the news articles to discover the correlations between events, and construct an event correlation evolution graph. Experiments test and validate our method.