Research On Automatic Semantic Annotation Method For The Internet Of Things Data

In the practical application of the Internet of Things(IoT),tens of thousands of information collection devices with different functions and forms are deployed into the network according to some specific network protocols.They exchanging information and communicating with each other continuously,and providing rich information resources for Internet of Things application services.However,at the same time,the massive,uncertain and heterogeneous nature of the raw data acquired by the front-end perception devices have greatly increasing the complexity of collaborative interaction,data fusion and analysis and reasoning among the resources of the Internet of Things.Meanwhile,it also brings great difficulties and limitations to the cross-domain sharing and reuse of data resources,which makes the application development of the Internet of Things fall into the dilemma between data isolation and the increasingly urgent demand for collaborative interaction.Therefore,this paper by using the advantages of semantic annotation technology in information description,representation and processing,with the purpose of shielding the heterogeneity and isolation of data,then making the users and machines better understand the meaning of data information generated by different devices,and excavating the deep value of various types of data,so as to improve the utilization rate of data and meet the complex and changeable needs of upper application.The main research work is described as follows:(1)An ontology hierarchical structure model of the Internet of Things domain is proposed,and the ontology modeling is realized.Study on the existing Internet of Things top-level domain ontology(SSN),targeted extract the relevant concept classes and object attribute relationships between classes that covered in the Internet of Things domain.On the basis,an IoT domain application ontology with special instance,and has the core content of "Device-Observation-condition" is constructed.Moreover,the ontology content is parsed to obtain the RDF triad statement for each ontology concept.(2)An automatic semantic annotation model for IoT data resources and a semantic data association model are proposed and constructed.The annotation model is based on the constructed application ontology,and we use the semantic similarity algorithm which combining word-form and WordNet meaning similarity computing method,as well as the automatic semantic annotation algorithm based on the thought of maximum similarity matching for RDF graph to conduct semantic information extraction and concept mapping of original IoT data resources,and then realize a complete service that from data acquisition to feedback data which contains semantic tags to users,so that to complete the explicit information annotation for data resources.At the same time,based on the RDF annotation document,the semantic association at the time level between related data resources is realized according to the data association model and certain association rules,so as to complete the implicit information annotation for data resources.Finally,it is proved by experiments that the semantic annotation model can effectively grant the IoT data resources a rich and clear standardized semantic annotation as well as an URI that can uniquely identify this information,meanwhile,the data semantic association result can improve the readability of the semantic annotation result to a certain extent.(3)Based on the automatic semantic annotation model and the semantic similarity mapping annotation algorithm,an automatic semantic annotation system for IoT data was designed and implemented.The core function of this system is to realize the automatic semantic annotation for the documents-oriented IoT data,which based on the application domain ontology of the Internet of Things.Moreover,this system still realizes the functions of semantic data information extraction,storage management and association visualization and so on.System test results show that the semantic annotation system developed in this paper can stably realize the automatic semantic annotation for data resource in the field of IoT,and has good performance and certain practical value.To a certain extent,this research has basically realized the transformation from the original data of the Internet of Things to the machine-understandable semantic data with rich grammatical structure.This work aims to obtain better data,and provide effective foundation support for upper-level semantic retrieval and reasoning applications,which has certain practical significance and application value.Besides,a practical application scenarios are introduced to analyze the feasibility of the annotation method,and the effectiveness of the proposed information processing method is verified.