Constructing Hierarchical Topic Knowledge Window for Chinese News Browsing and Retrieval

This paper proposes an event-based hierarchical knowledge structure for Chinese news display. We apply Hierarchical Topic Detection and Tracking technique to group news reports into clusters. Considering the extensibility of events, the knowledge structure is designed with robustness to merge similar reports and reorganize the structure automatically. This brings the merit of tracking associated follow-up event. To verify the applicable features, we design an assessing mechanism and call for participation for two weeks. There are 75 percent of participants expressing their high remarks on visualized knowledge expression. Eighty-two percent of participants consider the event heading is meaningful and acceptable. Over 90 percent of participants consider this study to be applicable.