RWTH Aachen University
In this paper, the authors present their work on the enrichment of the EventMedia dataset as provided by the DeRiVE data challenge with links to DBpedia. Their main contribution is an exploration into the use of domain knowledge in the matching process. As a starting point they take DBpedia spotlight, an off-the-shelf tool for matching textual resources to DBpedia. They present a bootstrap method to automatically derive the needed domain knowledge from an initial set of high confidence matches, and compare this to a baseline method without any domain knowledge, and an 'Oracle' method with perfect domain knowledge.