A Smart Framework to Detect Threats and Protect Data of IoT Based on Machine Learning

Abstract

The widespread use of IoT devices has introduced new challenges, particularly related to security and privacy threats such as unauthorized data access and device breaches. The high demand for these devices, coupled with the lack of strong security and privacy systems for user data, highlights the need for more effective approaches and solutions to address these issues. In this study, we propose a framework designed to audit, test, and detect potential vulnerabilities within IoT environments and their applications. The main components of the proposed framework include a machine-learning algorithm for data classification and attack detection. The framework also introduces two additional features calculated by the edge computing layer (fog layer) to enhance the accuracy of the classification algorithm. The algorithm uses multiple classification models to improve accuracy. To assess our classification algorithm’s efficiency, we implemented it using a real IoT dataset known as the TON_IoT dataset. The results demonstrated the accuracy, efficiency, and precision of our approach in detecting and mitigating potential vulnerabilities and threats like Tamper Data, Injection, Denial of Service, and Backdoor attacks. We achieved an accuracy rate of over 98%. Our findings offer valuable recommendations to enhance security and privacy within IoT systems, and we also explore emerging trends in these areas.

Keywords

• Internet of Things (IoT)
• attacks
• security
• deep learning
• fog computing
• privacy