A Novel Secure Data Processing Mechanism in IoT Using Deep Learning with Ontology

Security has become the biggest concern in recent technologies like the Internet of things (IoT) as it involves heterogeneous users and resources sharing sensitive data through cloud. These heterogeneous users and service providers have multifarious privacy and security requirements and need a common mechanism to share the same across the heterogeneous environments. Ontology proves to be an efficient means to handle heterogeneous environments for the following reasons: It provides simple means to share domain knowledge among entities, it is simple to reuse domain knowledge, and it facilitates convenient means to manage and manipulate domain entities and interrelationships among entities. The performance of ontologies is determined by their reasoning ability. In IoT, many devices are involved, and hence, multiple ontologies are to be involved. Still, most of the works mainly focus on only single ontology for reasoning. They lack considering reasoning involving multiple ontologies. This article proposed a deep learning method for associating various ontology rule bases and thus learning new inference rules and thereby providing efficient security in IoT applications. To verify the usefulness of the proposed work, it is realized in healthcare application and proves to achieve better security.