Exploiting the collective intelligence contained in Wikipedia to automatically describe the content of a document

The Wikipedia phenomenon very interesting from the point of view of the collective, social effort to produce a large, strongly interlinked body of knowledge. It also offers, for the first time in history, a general source of information coded in electronic form and freely available to anyone. As such, it can be used as a reference for tools aiming at mining semantic meaning from generic documents. In this paper, we propose a clustering-based method that exploits some of the implicit knowledge built into Wikipedia to refine and ameliorate existing approaches.