Dynamic Evaluation of Ontologies

• Published in: Procedia computer science 2015, Vol.73, p.16-23
• Main Authors: Lakel, Kheira, Bendella, Fatima
• Format: Article
• Language: English
• Subjects: Construction of ontology; multi-agent systems; ontology; ontology engineering; ontology learning; Semantic Web
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2015.12.043

The automatic construction of ontologies from texts is a topic of continued and open research, their construction requires both a study of human knowledge, methodologies and tools to retrieve the text content. As the content of these resources is dynamic, they can be thought of as finished products and refined, which remain stable when completed. The field of ontology construction needs to go towards more dynamic, more view of ontologies is to increase intelligence in many applications such as information retrieval, semantic indexing and semantic annotation. Ontologies are software modules whose development is based on the same principles as those applied in software engineering. There are several approaches for evaluating ontologies, some are based on learning methods from the corpus, using the networks head-Expansion or other semantic networks for identify concepts and relationships. But the automation of ontology construction, actually, is a scientific lock for many applications. In this paper, we propose an approach that combines these tools to improve the process of automatic co-construction of ontologies from a corporus. DEO (Dynamic Evaluation of Ontologies) is a system dedicated to ontology construction from texts using a cooperative learning based on a multi agents structure. It uses the mechanism of extraction of relations Dynamo and more it uses the terminology WordNet1.2 to identify concepts, relationships and a storage module to save the changes of the ontologist in order to be reused.