Semi-automatic metadata annotation of Web of Things with knowledge base

Integrating semantic technology to the Web of Things (WoT) facilitates the creation of a networked knowledge infrastructure with more interoperable data from both physical and cyber world. However, most of current solutions relies on manual methods based on domain-specific ontology, which is only suitable for domain expert and developers but not compatible with large-scale knowledge construction. This paper proposes a semi-automatic annotation framework for the metadata representation of WoT resource. This framework is based on a probabilistic graphical model to collectively infer entities, classes and relations from schematic WoT resources mapping to domain independent knowledge base. A proof-of-concept implementation and performance evaluation are illustrated to show a feasibility of our method.