Hybrid Firefly and Particle Swarm Optimization for XGBoost Parameter Tuning Problem: Network Intrusion Detection

This research aimed was to optimize hyperparameters in NIDS using the XGBoost model by employing a hybrid metaheuristic approach, mixing Firefly Algorithm (FA) and Particle Swarm Optimization (PSO). To cope with the rapidly rising volume and complexity of cyber threats, traditional intrusion detection techniques suffer from a number of problems such as high false positive rates and low-level performance in accurate threat identification. This study shows a lack of detailed optimization methods and raises the urgent calls for novel approaches to improve machine learning models performance on IDS. The methodology was based on an experimental design using benchmark datasets, specifically NSL-KDD datasets, to evaluate the performance of the newly proposed hybrid optimization approach. The study puts emphasis on the technique of parameter tuning and its important impact on higher detection accuracy while lowering costs. The results show that the FA-PSO-XGBoost hybrid model compared favorably to the baseline or traditional optimization approaches while achieving a remarkably high detection accuracy of 0.9988, indicating it may serve well in a dynamic network environment in real-time. The result shows the promise of new innovative optimized approaches and offers insight for future NIDS research in cybersecurity. The study concludes with proposals for continued study and implementation using optimized machine techniques for network security against evolving cyber threats.