The neofonie NERD system at the ERD challenge 2014.

ABSTRACTThis paper describes Neofonie NERD, our Named Entity Recognition and Disambiguation system submitted to the ERD Challenge 2014. The system uses a vector space model approach for disambiguation, based on the link structure of Freebase, in combination with precomputed statistical measures from Wikipedia and Freebase. It was originally developed for the German language and has now been adapted for English. We achieved 70.0% F1-score in the final evaluation, which is 5.7 percent points above the average of all participating teams.