Hacia la identificación de relaciones de hiponimia/hiperonimia en Internet

this paper presents an approach to the automatic extraction of hyponyms and hyperonyms. in particular, it proposes an information extraction method that is specially suited for identifying pairs of hyponym-hyperonym by using a set of spanish lexical patterns. it also proposes a symmetric weighting scheme of pairs/patterns whose goal is to enhance the confidence of the extraction method. the effectiveness of the proposed approach was evaluated by extracting hyponyms from a given vocabulary of hyperonyms. results show the usefulness of the proposed extraction method as well as the relevance of the pairs/patterns weighting scheme.