查看文章

Graph neural networks (GNN) are effective methods for finding solutions for complex data structures in different data domains. In this study, we propose a GNN based approach to identify the source node which can generate data in an internet of things (IoT) network. We generate an IoT network and represent as a graph network individual nodes as nodes and links as edges. Each node is assigned different data as a node feature and assigned a unique label for each of the nodes. The proposed IoT network is simulated, and experiment results show the proposed model can achieve more than 90% accuracy in classifying nodes.