Request Paper

Abstract
Cybersecurity threats, including the increase of sophisticated supply chain attacks, continue to escalate in complexity and volume, necessitating the development of robust and efficient intrusion detection systems (IDS). This study presents an enhanced intrusion detection model that combines Binary Particle Swarm Optimization (BPSO) for feature selection with XGBoost for classification. By leveraging BPSO’s optimization capabilities and XGBoost’s ensemble learning strengths, the proposed BPSO-XGBoost model demonstrates superior performance in identifying malicious activities in network traffic. The methodology is validated using real-world cybersecurity datasets, achieving significant improvements in metrics such as sensitivity, F1 score, and AUC-ROC compared to traditional machine learning models. These results emphasize the model’s potential to strengthen IDS frameworks and mitigate advanced cyber threats effectively. Index Terms—Cybersecurity, Intrusion Detection Systems, Machine Learning, Feature Selection, BPSO-XGBoost.