Towards aligning IoT data with domain-specific ontologies through Semantic Web technologies and NLP

Internet of Things (IoT) data has the potential to be utilized in many domain-specific applications to enable smart sensing in areas that were not initially covered during the conceptualization phase of these applications. Typically, data collected in IoT scenarios serve a specific purpose and follow heterogeneous data models and domain-specific ontologies. Therefore, IoT data could not easily be integrated into domain-specific applications, as it requires ontology alignment of diverse data models with the end application. This poses a big challenge to semantic interoperability during the integration of IoT data into a pre-established system. In this line, the alignment process is cumbersome and challenging for an ontology engineer, since it requires a manual review of the relevant ontologies that could be aligned with the IoT data. Additionally, before aligning each term used in the IoT data with the concepts defined in the domain-specific ontologies, all similar/related terms in the given ontologies must be considered. In this paper, we propose a solution that supports the alignment process by utilizing semantic web technologies and Natural Language Processing (NLP). Our novel solution proposes an NLP-based term alignment with a similarity score that supports identifying the relevant terms used in IoT data and ontologies and stores the similarity scores among terms based on different similarity algorithms. We showcase our solution by aligning IoT sensor data with the water and IoT domain ontologies.