| Abstrakt: |
This research explores using a Federated Random Forest (FRF) model with feature selection techniques to improve collaborative intrusion detection in Internet of Things (IoT) settings. In the world of IoT security, where devices have limited resources and privacy is crucial, this study aims to overcome challenges like resource constraints and data privacy. The FRF model, building on the effectiveness of Random Forest (RF), is enhanced with feature selection methods to improve intrusion detection accuracy while managing dimensionality issues. By allowing devices to work together on intrusion detection without revealing sensitive data, this research aims to establish a stronger and privacy-conscious IoT security framework. The study also investigates the performance of the FRF model adapts to different types of IoT devices, ensuring its effectiveness in diverse environments. The findings aim to offer valuable insights and practical solutions in the evolving field of collaborative intrusion detection in IoT, using the CIC IoT Dataset 2023. Based on the experiment result, the proposed model achieves an outstanding ± 99.68% accuracy, precision, recall, and F1-Score, indicating remarkable correctness and balance in its predictions for intrusions. This exceptional performance underscores the model's reliability and effectiveness across all evaluation metrics. [ABSTRACT FROM AUTHOR] |
