Abstract
Today we read the news from a variety of sources in varying depth. A single article is generally part of a long running story often requiring context to fully comprehend it. Jour- nalists, historians and social scientists prefer exploring news archive in order to study the context of a news article. Typi- cal news archive exploration systems complement result lists with temporal and topical insights solely from the underly- ing document collection to contextualize which allows users to intelligently refine the intent. However these insights are devoid of user impressions and are only reliable if the un- derlying archive is replete and coherent. For web crawled news archives that are often incoherent we propose an ap- proach to insight mining for exploration using Wikipedia’s construction dynamics. In this paper we demonstrate the advantages of using Wikipedia for news exploration with a prototype system that builds on top of pevious work in news exploration.
Links and resources
Tags
community
@j_singh's tags highlighted