Ensemble technique of intrusion detection for IoT-edge platform

Area of Science:

• Computer Science
• Cybersecurity
• Network Engineering

Background:

• Internet of Things (IoT) technology adoption presents significant security challenges due to device resource limitations.
• Effective intrusion detection systems are critical for mitigating cyberattacks in interconnected IoT ecosystems.
• Existing security measures often struggle to adequately address the unique vulnerabilities of IoT networks.

Purpose of the Study:

• To develop and evaluate a robust two-stage intrusion detection procedure for Internet of Things (IoT) environments.
• To enhance the accuracy and efficiency of identifying and classifying network intrusions in IoT data traffic.
• To provide a reliable security framework that addresses the overlooked security concerns in resource-constrained IoT devices.

Main Methods:

• A two-stage intrusion detection methodology was proposed, beginning with an Extra Tree (E-Tree) binary classifier.
• The second stage employed an Ensemble Technique (ET) integrating E-Tree, Deep Neural Network (DNN), and Random Forest (RF) for invasive event analysis.
• Performance was rigorously assessed using diverse datasets: Bot-IoT, CICIDS2018, NSL-KDD, and IoTID20.

Main Results:

• The proposed two-stage intrusion detection strategy demonstrated superior performance compared to existing machine learning methods.
• The system achieved enhanced statistical measures, including accuracy, normalized accuracy, recall, and stability.
• Experimental validation confirmed the effectiveness of the integrated E-Tree, DNN, and RF ensemble approach in detecting IoT intrusions.

Conclusions:

• The developed two-stage intrusion detection system offers a significant advancement in securing Internet of Things (IoT) networks.
• The combination of Extra Tree and Ensemble Techniques provides a powerful and accurate method for identifying sophisticated cyber threats.
• This research highlights the potential of advanced machine learning models in addressing critical security gaps in the evolving IoT landscape.