Bridging Signals and Human Intelligence - Log Mining-Driven and Meta Model-Guided Ontology Population in Large-Scale IoT
Sprache des Titels:
Proceeding of the 15th International Conference on Knowledge Science, Engineering and Management (KSEM), Singapore, August 6?8, 2022
Large-scale Internet-of-Things (IoT) environments such as Intelligent Transportation Systems are facing tremendous challenges wrt. monitoring their operational technology (OT) not least due to its inherent heterogeneous and evolutionary nature. This situation is often aggravated by the lack of machine-interpretable information about the interdependencies between OT objects in terms of ?semantic relationships?, thus considerably impeding the detection of root causes of cross-system errors or interrelated impacts. Therefore, we propose a novel hybrid approach for identifying semantic relationships based on both, mined functional correlations between OT objects based on log files and domain knowledge in terms of an IoT meta model. For this, we firstly contribute a systematic discussion of associated challenges faced in large-scale IoT environments, secondly, we put forward an IoT meta model based on both, industry standards and academic proposals, and finally, we employ this meta model as guidance and target template for the automatic population of semantic relationships into an OT ontology.