Context-aware user-driven news recommendation

Abstract

Recommender systems match available contents with users' contexts and interests. With linked data knowledge bases we can build recommender systems where user interests, their context and available contents are modeled in terms of real world entities. In this demo paper we will describe existing academic news recommender systems and the Smartmedia prototype in particular. This prototype shows how we can combine available technologies like semantics, natural language processing and information retrieval to construct personalized and location aware recommendations on a continuous stream of news information.