OBSemE: An ontology-based semantic metadata extraction system for learning objects

Abstract

In this paper we describe OBSemE an ontology-based semantic metadata extraction system which implement our approach dedicated to the automatic extraction of semantic metadata for learning objects. The process of semantic metadata extraction is based on ontology metadata information extraction method's principles. This choice is due to the advantages of the use of ontologies. The input of our system is a set of LOM metadata elements respecting three requirements. The first requirement is that each chosen LOM data element must describe the educational content of the learning object. The second requirement is that the LOM data element must be required by most of the widely used LOM application profiles. The third requirement is that the LOM data element has to be mostly fulfilled by the learning object's authors in practice. The output of our system is a set of semantic metadata describing the learning object content. We have designed an RDF schema to encode semantic metadata with a computable formalism. We provide also some experimental results as a proof of the feasibility of our approach and the quality of our implementation.