Redada: Mining Knowledge out of Italian Business News Items

The paper describes Redada, a fully-implemented ontology-based relation extraction system that integrates classical NLP techniques with expert knowledge expressed by ontologies. Experimenting with the system for the task of supporting law enforcement and intelligence activities against money laundering and corruption, exploring information out of Italian newspapers to help in the identification of networks of activities, both legal and illegal, we evaluate the proposed system against a standard baseline approach, and we present our first results.