Related Experiment Videos
Explainable Intrusion and Anomaly Detection for IoT Sensor Networks Using Hybrid Feature Selection and Deep
Usman Ahmed1, Sadiq Muhammad2, Jaeyoung Choi2
1School of Software, Northwestern Polytechnical University, Changan Campus, Xi'an 710072, China.
Sensors (Basel, Switzerland)
|July 28, 2026
Summary
This study introduces HDATL-XAI, an intrusion detection system (IDS) for Internet of Things (IoT) security. It uses hybrid feature selection, anomaly detection, and explainable AI to enhance threat identification and transparency in IoT networks.
Area of Science:
- Cybersecurity
- Artificial Intelligence
- Network Security
Background:
- Internet of Things (IoT) environments face unique security challenges due to their distributed and resource-constrained nature.
- Developing robust and transparent Intrusion Detection System (IDS) solutions is crucial for securing IoT ecosystems.
- Existing IDS solutions may struggle with the complexity and scale of modern IoT networks.
Purpose of the Study:
- To present a specialized system for IoT environments that enhances security measures and event transparency.
- To integrate hybrid feature selection, anomaly detection, classification, and explainability techniques for improved IoT security.
- To provide a trustworthy and transparent framework for identifying and classifying cyber threats in IoT networks.
Main Methods:
- Hybrid feature selection combining Random Forest (RF) and SelectKBest to reduce computational overhead.
- Deep Autoencoders (DAEs) for anomaly detection, identifying traffic deviating from normal behavior.
- Feedforward Neural Networks (FNNs) for accurate and efficient classification of anomalous data.
- Explainability tools like Shapley Additive Explanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) for model transparency.
Main Results:
- The proposed HDATL-XAI framework effectively identifies critical cyber threats in IoT environments.
- The system demonstrates high accuracy in detecting anomalous traffic and classifying threats.
- Explainability techniques provide insights into model decisions, enhancing trust and transparency.
Conclusions:
- The HDATL-XAI framework offers a comprehensive approach to IoT security, integrating advanced AI techniques.
- The system enhances both the security and trustworthiness of IoT networks through transparency.
- This framework serves as an essential tool for protecting vulnerable IoT infrastructure against cyber threats.