Building Domain Ontologies from Text for Educational Purposes (revised July 2008)

This paper presents a semi-automatic framework that aims to produce domain concept maps from text and then to derive domain ontologies from these concept maps. This methodology particularly targets the eLearning and AIED (Artificial Intelligence in Education) communities as they need such structures to sustain the production of eLearning resources tailored to learners’ needs. This paper details the steps to transform textual resources, particularly textual learning objects (LOs), into domain concept maps and it explains how this abstract structure is transformed into a formal domain ontology. A methodology is also presented to evaluate the results of ontology learning. The paper shows how such structures (domain concept maps and formal ontologies) make it possible to bridge the gap between eLearning and Intelligent Tutoring Systems by providing a common domain model.