Leveraging Ensemble Methods for Robust Attack Detection in Consumer IoT Environments

Consumer Internet of Things (CIoT) enables seamless communication between smart devices but lacks robust security measures. This vulnerability exposes CIoT systems to cyberattacks like Distributed Denial of Service (DDoS) and unauthorized access. Major studies indicate that traditional ML algorithms struggle with accuracy in attack detection. To address this, this paper proposes a voting classifier combining Decision Tree (DT), Random Forest (RF), KNN, and SVM to enhance cyberattack detection in complex CIoT environments. This approach reduces the limitations of individual methods and enhances robustness by utilizing their advanced capabilities. The experimental results show 96.23% detection accuracy with this novel approach outperforming the individual models and thus gives a promising solution for environment.