HSGAN-IoT: A hierarchical semi-supervised generative adversarial networks for IoT device classification
Yanliang Jin,Jiahao Zhou,Yuan Gao
DOI: https://doi.org/10.1016/j.comnet.2024.110299
IF: 5.493
2024-02-01
Computer Networks
Abstract:In recent years, IoT device classification has become a highly focused issue, because it can achieve network performance optimization, security threat detection, application scenario analysis and device performance evaluation. Existing researches mainly focus on two types of methods: machine learning (ML) and deep learning (DL). However, with the continuous emergence of new devices, ML methods require experts to constantly update and design high-quality features, while DL methods require capturing new data from raw traffic and manually labeling it. These processes are time-consuming and error prone. To solve the problems above, this paper proposes a hierarchical semi-supervised generative adversarial networks for IoT device classification called HSGAN-IoT. HSGAN-IoT proposes a new structure and joint loss to make adversarial training focus on classification tasks instead of generating tasks. Compared to traditional models, HSGAN-IoT classifies a device by three attributes: type, function, and manufacturer. This enables it to classify unknown devices which have not appeared during training. Moreover, HSGAN-IoT avoids redundant calculations through its hierarchical classification structure, which filters out IoT devices from raw traffic data in the first layer and performs further classification in the second layer. This paper also establishes a fusion dataset to test the classification performance of HSGAN-IoT. The results show that HSGAN-IoT is superior to the state-of-the-art researches, with accuracy rates of 99.56%, 96.97%, and 97.86% on three classification tasks, respectively.
computer science, information systems,telecommunications,engineering, electrical & electronic, hardware & architecture