Abstract:

Systematically and semantically modeling the 5W1H (What, Who, When, Where, Why and How) elements of a news article is a fundamental step towards automatic understanding of Chinese news events. In this paper, we propose an ontologybased event model, NOEM, which can semantically describe the 5W1H elements of an event and further organize them in the form of an ontology. The proposed model defines the concepts of entities (time, person, location, organization, etc.), events and their relationship in order to capture the different natures of news events, including temporal, spatial, causal aspects and so on. Additionally the wellorganized event elements are essential for building an event knowledge base which can significantly facilitate online news browsing and managing in various applications. A comparison to the existing event models and an empirical case study show that NOEM can effectively model the semantic elements of news events and their relationship; and has a strong ability to represent the knowledge facts and easily adapt to new domains.

Key words: 5W1H;ontology;event model;ontology population