A Study on News Ontology Construction using External Knowledge and Collective Intelligence

Abstract

News sites such as Google News allow Internet users to look up news articles. As a more advanced service, we are attempting to develop a "Multimedia News Aggregator" framework, which retrieves groupings of news articles related to the user's interests with reconstruction in conjunction with news video. To retrieve and reconstruct suitable news articles related to the user 's requirements, we must define various news elements, such as "Abstract," "Cause," and "Effect," which compose news events, and define the concepts that compose the news elements. This paper proposes a news ontology construction technique that expresses the structure of such news events. By using Wikipedia and WordNet as sources of external knowledge, we generalize the combinations of element of news events, and the concepts of words in the elements.