Leveraging XGBoost based GBM for Proactive Detection of Man-in-the-Middle Cyber Attacks
Shivanand V. Manjaragi, Deepak Asrani, Arvind Kumar, R Sandesh, Meenakshi Maindola, A. K. M. Sai · 2024
Man-in-the-Middle (MitM) cyber-attacks provide a substantial risk to network security, since they allow hostile individuals to intercept and alter communications between two parties. Conventional detection approaches frequently fail to recognize these complex threats, requiring the creation of more resilient and precise solutions. This study introduces a model called XGBoost-based Gradient Boosting Machine (GBM) to identify MitM attacks in advance. The system employs extensive datasets from Kaggle, namely CICIDS 2017 and UNSW-NB15, for the purpose of training and evaluating the model. By carefully preparing the data, selecting relevant features, and doing thorough cross-validation with hyperparameter optimization, the suggested model attains exceptional performance metrics. The assessment findings indicate an accuracy of 0.98, precision of 0.97, recall of 0.96, F1-Score of 0.965, and a remarkable ROC-AUC score of 0.99. Our model regularly outperforms existing technologies like as Random Forest and Traditional GBM in identifying MitM attacks, achieving higher performance metrics across the board. This research demonstrates the effectiveness of the XGBoost-based GBM model in improving cybersecurity defenses and suggests that it might serve as a standard for future advancements in cyber-attack detection systems. The suggested system’s exceptional accuracy and resilience make it an invaluable tool for proactively detecting and reducing MitM attacks.