Evaluation of Selected Stacked Ensemble Models for the Optimal Multi-class Cyber-Attacks Detection

The significant rise in the frequency and sophistication of cyber-attacks their diversity necessitated various researchers to develop strong effective approaches address recurring cyber threat challenges. This study evaluated performance three selected meta-learning models for optimal multi-class detection using University New South Wales 2015 Network benchmark (UNSW-NB15) Intrusion Dataset. results this show confirm ability base models; Naive Bayes, C4.5 Decision Tree, K-Nearest Neighbor solving problems. It further affirms knack duo feature selection techniques stacked ensemble learning optimize ML models' performances. stacking predictions information gain with Model Tree meta-algorithm recorded most improved accuracy Mattew's correlation Coefficient than Multiple Trees (MMT) Multi Response Linear regression (MLR) Meta algorithms.