IOS Press Disclose High Quality Structured Data with

Abstract. The advent of Wikipedia as the best digital representation of cross-domain knowledge is now a reality. The DBpedia project aims at converting Wikipedia content into structured intelligence through the Linked Open Data paradigm and currently holds a vital role in the growth of the Web of Data as a multilingual interlinking kernel. However, its main classification system relies on a time-consuming manual procedure that aims at mapping Wikipedia infobox data to a common ontology. We present Airpedia, a tool that automatically generates class and property mappings for any DBpedia language chapter. We support our findings with the deployment of the Swedish, Ukrainian, and Esperanto DBpedias. Evaluations demonstrate that Airpedia is not only comparable to humans in terms of precision, but also provides a recall leap, with special regard to entities lacking infoboxes