Research on event-based semantic annotation of Chinese

Nowadays most of event-based corpuses only have English version. Existing event tagging methods are inappropriate for processing Chinese texts, as they are insufficient in terms of expression for event-relations and function words. Chinese is a flexible language, which does not follow strict grammar and contains various function words. To make Chinese text possible for computer analysis, sentences should be processed from event perspectives. This paper Chinese sentence are divided into event denoters, event-elements and nonevent-elements, semantically tags each of components and classifies event non-taxonomic relationships by time and logic. Furthermore, we give a new XML approach to represent the tagging system, which laid groundwork for construction of large-scale, event-based Chinese corpuses.