A model of an ontology oriented threat detection system (OOTDS)

The paper discusses the design of an ontology oriented threat detection system (OOTDS) for an environment of objects with properties and relations described by random variables in a PR-OWL first order Bayesian logic ontology. The environment is monitored by a set of monitors continuously providing the OOTDS with information about current events, which reflect changes in the environment. Thanks to the PR-OWL ontology and MEBN reasoning process the system can interpret the arising events as premises of possible threats. The users can indicate environment states that they consider as threats and define various notification ways for different threat types. We give a short introduction to MEBN logic followed by an exemplary OOTDS usage scenario and finish considerations with remarks regarding future work.