Enterprise ontology learning for heterogeneous graphs extraction

Abstract

In the enterprise context, people need to visualize different types of interactions between heterogeneous objects in order to make the right decision. Therefore, we have proposed, in previous works, an approach of enterprise object graphs extraction which describes these interactions. One of the steps involved in this approach consists in identifying automatically the enterprise objects. Since the enterprise ontology has been used for describing enterprise objects and processes, we propose to integrate it in this process. The main contribution of this work is to propose an approach for enterprise ontology learning coping with both generic and specific aspects of enterprise information. It is three-folded: First, general enterprise ontology is semi-automatically built in order to represent general aspects. Second, ontology learning method is applied to enrich and populate this latter with specific aspects. Finally, the resulting ontology is used to identify objects in the graph extraction process. © 2012 Springer-Verlag.