Towards Probabilistic Ontology based on Bayesian Networks

Abstract: The increase and diversification of information has created new user requirements. The problems of representation, processing, analysis and reasoning on the information, especially the uncertain information, are still an important research topic. We try to show in this paper how this issue is crucial. It is essential to propose new approaches to formally represent uncertain information to help machines understand and infer new knowledge. At first, we will start by defining the Bayesian networks, the way of it construction. Then will then focus on probabilistic ontologies and their relationship with Bayesian Networks. Then, we will show how we can integrate a Bayesian network into a probabilistic ontology citing the translation rules which is the first step of the process of integration. Finally we conclude with perspectives of our work