作者: Andrea Giovanni Nuzzolese , Aldo Gangemi , Valentina Presutti , Paolo Ciancarini
DOI:
关键词:
摘要: DBpedia contains millions of untyped entities, either if we consider the native ontology, or Yago plus WordNet. Is it possible to automatically classify those entities? Based on previous work wikilink invariances, wondered wikilinks convey a knowledge rich enough for their classication. In this paper give three contributions. Concerning link structure, describe some measurements and notice both problems (e.g. bias that could be induced by incomplete ontological coverage ontology), potentials existing in current type coverage. classication, present two techniques exploit wikilinks, one based induction from machine learning techniques, other abduction. Finally, discuss limited results which conrmed our fears expressed description general gures measurement. We also suggest new directions entity classication taken.