A framework for anomaly detection and classification in Multiple IoT scenarios

The investigation of anomalies is an important element in many scientific research fields. In recent years, this activity has been also extended to social networking and social internetworking, where different networks interact with each other. In these research fields, we have recently witnessed an important evolution because, beside networks of people, networks of things are becoming increasingly common. IoT and Multiple IoT scenarios are thus more and more studied. This paper represents a first attempt to investigate anomalies in a Multiple IoT scenario (MIoT). First, we propose a new methodological framework that can make future investigations in this research field easier, coherent, and uniform. Then, in the context of anomaly detection in an MIoT, we define the so-called "forward problem" and "inverse problem". The definition of these problems allows the investigation of how anomalies depend on inter-node distances, the size of IoT networks, and the degree centrality and closeness centrality of anomalous nodes. The approach proposed herein is applied to a smart city scenario, which is a typical MIoT. Here, data coming from sensors and social networks can boost smart lighting in order to provide citizens with a smart and safe environment. (C) 2020 Elsevier B.V. All rights reserved.