EMMBTT: A Novel Event Evolution Model Based on TFxIEF and TDC in Tracking News Streams

With the popularity of the Internet, online news media are pouring numerous of news reports into the Internet every day. People get lost in the information explosion. Although the existing methods are able to extract news reports according to key words, and aggregate news reports into stories or events, they just list the related reports or events in order. Moreover, they are unable to provide the evolution relationships between events within a topic, thus people hardly capture the events development vein. In order to mine the underlying evolution relationships between events within the topic, we propose a novel event evolution Model in this paper. This model utilizes TFIEF and Temporal Distance Cost factor (TDC) to model the event evolution relationships. we construct event evolution relationships map to show the events development vein. The experimental evaluation on real dataset show that our technique precedes the baseline technique.