Recommender systems meet Linked Open Data

Abstract

Information overload is a problem we daily experience when accessing information channels such as a Web site, a mobile application or even our set-top box. There is a clear need for applications able to guide users through an apparently chaotic information space thus filtering, in a personalized way, only those elements that may result of interest to them. Together with the transformation of the Web from a distributed and hyperlinked repository of documents to a distributed repository of structured knowledge, in the last years, a new generation of recommendation engines has emerged. As of today, we have a huge amount of RDF data published as Linked Open Data (LOD) and available via a SPARQL endpoint and the number of applications able to exploit the knowledge they encoe is growing consistently. Among these new applications and services, recommender systems are gaining positions in the LOD arena.