Statistical Profiles of Words for Ontology Enrichment
Andreas Faatz, Cornelia Seeberg, Ralf Steinmetz · Advances in intelligent and soft computing · 2002
The following paper focuses on an enrichment method for ontologies. We define similarities of possible new concepts and base the similarity and dissimilarity of concepts on the usage statistics in large corpora. The method is soft in the sense, that we define a semantically motivated heuristics for the influence of different linguistic properties influencing the similarity definition.