Related Experiment Videos
Hyperparameter optimization of XGBoost and hybrid CnnSVM for cyber threat detection using modified Harris hawks
Haitham Elwahsh1,2, Ali Bakhiet3, Tarek Khalifa4
1Faculty of Information Technology, Applied Science University, Amman, Jordan.
Peerj. Computer Science
|September 24, 2025
Summary
This study introduces a new cyber threat detection framework for smart microgrids using Harris Hawks Optimization (HHO) to boost machine learning models. The optimized models achieved over 99.97% accuracy in detecting sophisticated cyber attacks.
Area of Science:
- Cybersecurity
- Artificial Intelligence
- Smart Grid Technology
Background:
- Smart microgrids face increasingly complex cyber threats.
- Existing detection methods often fail to optimize effectively or handle imbalanced datasets.
Purpose of the Study:
- To propose a novel framework for enhanced cyber threat detection in smart microgrids.
- To address limitations in current optimization techniques and class imbalance issues.
Main Methods:
- Integration of Harris Hawks Optimization (HHO) for hyperparameter tuning.
- Utilizing Extreme Gradient Boosting (XGBoost) and a hybrid Convolutional Neural Network-Support Vector Machine (CNN-SVM).
- Employing RandomOverSampler to manage class imbalance in datasets.
Main Results:
- Achieved high detection accuracies of 99.97% and 99.99% on DDoS botnet and KDD CUP99 datasets.
- Demonstrated improved Area Under Curve (AUC) metrics, indicating superior performance.
- Successfully captured complex nonlinearities within the data.
Conclusions:
- The HHO-optimized framework offers a robust and scalable solution for automated threat detection.
- This approach significantly enhances the security of critical infrastructures against advanced cyber threats.
