Presents a rigorous cross-dataset benchmark of Gradient Boosted Decision Tree models for network intrusion detection, demonstrating the effectiveness of Bayesian hyperparameter optimisation in improving detection performance, robustness, and generalisation across diverse intrusion datasets.
