Ontology-based Automatic Semantic Annotation Method for IoT Data Resources

Abstract

The Internet of Things (IoT) system contains tens of thousands of front-end sensing devices with different functions and various patterns of manifestation. Therefore, the raw IoT data acquired by those devices was born with the feature of multi-source and heterogeneous, which leads to the difficulties and challenges for collaborative interaction and data fusion between IoT resources. Therefore, how to shield the isolation and uncertainty of IoT data, realize the cross-domain sharing and reusing of data resources are the key issues of IoT data processing field. By introducing the semantic technology, this paper constructs domain ontology and automatic semantic annotation model for IoT data, and then based on the semantic similarity mapping algorithm, adds semantic tags for the raw IoT data. Moreover, an association model with the core content of “data resource-observation time” is constructed, so that to obtain high-quality IoT data information further. In the end, under the background of a smart greenhouse instance, by calculating the values of the three evaluation indicators of precision rate P, recall rate R and F value to analyze the performance of the annotation model, the availability of the proposed method is verified. In addition, the implementation of annotation semantic tagging system proves that the proposed method can quickly and accurately complete IoT data annotation, which further presented the good performance and practical value of the annotation method. Which would lay a theoretical foundation and technical support for the intelligent decision-making and inferential retrieval of the Internet of Things system.