Construction incrémental d'une ontologie par analyse du texte et de la structure des documents

Abstract

BIOTIM is a project to design generic methods for the automatic analysis of large amounts of texts and images in order to acquire a common semantic layer and, building upon this initial result, to develop generic methods for a multi-modal examination of the structured data obtained. As part of this project we present an experiment aimed at improving the knowledge acquisition process by exploiting simultaneously both the structure and textual content of documents. In particular, we show how to build an intermediary ontology in the field of exotic flowers (orchidae family) and how this intermediary ontology represents an incremental step in in the building of a complete domain ontology. © Lavoisier. Tous droits réservés pour tous pays.