Reseatch On Lightweight Sensor Otology Mapping In Semantic Web Of Things

Semantic Web of Things (SWoT) is the solutions of the inner contradictory Internet of things (IoT). Introducing the concept of ontology into the IoT to realize the semantics of knowledge and collaboration of IoT is one of the core contents of S WoT. With the development of SWoT, the number of sensor ontologies is increased dramatically. But sensor ontologies built by different experts and fields are not consistent for no unified standard. In order to achieve the semantic collaboration between sensor ontologies, there is a need to build mappings of sensor ontologies.At present, ontologies are mainly divided into two categories:descriptive ontology and directory ontology. Many sensor ontologies of SWoT are lightweight ontologies which are an important branch of directory ontology. In the process of mapping between two sensor ontologies, considering the directory structure characteristics of ontologies, we find that using existing S-Match ontology mapping method in the mapping of lightweight sensor ontologies is suitable. Although the traditional S-Match method solves many problems, but this method has some problems:(1)Due to the calls of the SAT reasoning machine frequently, semantic mapping requires a lot of time.(2)The low recall rate caused by the lack of background knowledge.In order to overcome problem (1), this paper proposes a lightweight sensor ontology mapping framework, CMSB_LSOMF, based on the center mapping set to improve the S-Match. First of all, the paper puts forward the definitions of center mapping set and redundant mapping set. Center mapping set is a subset of initial mapping sets and reduces the redundant mapping of them. Then the program get the center mapping set and redundant set. Finally, conflicting mappings existing in mapping set are processed. In order to overcome problem (2), WordNet is used to enrich background knowledge and enlarge semantic information, which makes CMSB_LSOMF have a good performance in recall and precision.Finally, this paper conduct an experiment by using the JAVA language, Jena and Protege. Experiment shows that compared with the original S-Match method, CMSB_LSOMF has some improvement in the time efficiency and has some promotion in recall and precision by reducing the amount of mapping element and the number of SAT reasoning machine calls.