IoT Security using Machine Learning: Applications and Difficulties based on Forthcoming Directions
Vijay Anant Athavale, Atiya Sultana, A Khadeeja Bilquees, M. Fathima, Muhammed Basheer, Kavita Khatana · 2023
The Internet of Things is becoming mainstream in all computer science fields. In the near future, the Internet of Things (IoT) will affect the economy, enterprises, and society. However, the cross-cutting nature of transdisciplinary components and IoT systems utilized in such projects has revealed new security concerns. Nodes in an Internet of Things network generally lack resources, making them ideal targets for hackers. Internet of Things devices have basic security weaknesses, therefore authentication, encryption, application security, and access networks are worthless. This architecture uses NFV and SDN enablers to mitigate system vulnerabilities. This artificial intelligence framework uses anomaly-based models to detect IoT system intrusions and network trends using a monitoring agent and an AI-based response agent. ML approaches are needed to change IoT system protection from secure communication to intelligent security solutions. Test results prove the recommended strategy can work. Data mining may be a cost-effective strategy to find high-performance attacks. Specifically, this may detect attacks. This is because it emphasizes dispersion data. We used a single-class support vector machine to test our anomaly-based IDS for the Internet of Things. A real intelligent building environment was used for the assessment. Abnormalities were correctly identified 99.71 percent of the time. An idea’s practicability is investigated to find existing answers and encourage study into unsolved challenges.