To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
A Hybrid Metaheuristic Framework for High-Dimensional Cyber-Infiltration Detection & Mitigation |
|
Author |
Nazar Faried Yousif Mohamed, Ashraf Gasim Elsid Abdalla, AbuBakr Obeid Mosa Mohammed, Hashim Elshafie, Rania Ali Elkhidir Ali, Hashim Albasheer, Omer Elsier Tayfour, |
| Citation |
Vol. 26 No. 9 pp. 134-145 |
|
Abstract |
Distributed Denial of Service (DDoS) attacks are among the most severe entities that threaten network infrastructures in contemporary times because of their growing frequency and complexity. This paper presents a scalable and smart DDoS detection model that builds on machine learning and real-time analytics to improve detection effectiveness and response performance. The approach suggested is a hybrid ensemble model, a combination of the Random Forest and XGBoost with soft voting to deliver high classification accuracy and strength. In order to enhance the computational efficiency, there is the feature selection mechanism that is founded on the Bat Algorithm, which minimizes the data dimensionality without eliminating the important discriminative features. In addition, a real-time analytics engine is created to allow the analysis of traffic and quick decision-making. A lightweight integrating layer, which is a JSON-based one, is meant to enable smooth interaction between the parts of the system, such as the detection engine, the visualization dashboard, and the alerting module. An automated response and mitigation system is also integrated into the system and is able to perform immediate actions in response, like blocking and filtering of traffic. The experimental findings of the CIC-DDoS2019 dataset show that the proposed framework yields a training accuracy of 99.99% and a testing accuracy of 99.94% with a precision of 99.97, a recall of 99.95, and an F1-score of 99.96. Also, the model has a high specificity of 99.89 and a low false positive rate (FPR) of 0.11, demonstrating good detection and high accuracy of generalization. These findings provide evidence of the suitability of the suggested system in practical use in network security settings. |
|
Keywords |
Intrusion Detection System, Hybrid Wrapper (Bat?RF) AlgorithmEnsemble Learning, XGBoost, DDoS |
|
URL |
