EFSMDCD: Method for Detecting and Categorzing Dos and DDoS Attacks using Ensemble Feature Selection

The issue of cyber security has gained significant global attention over the networks (Cloud Computing, Internet of Things(IoT), and Software Defined Networks (SDN)). To identifying malicious individuals and actions, intrusion detection systems (IDS) are essential for safeguarding networks that are connected. An intrusion detection system (IDS) that uses machine learning (ML) behavior analysis may successfully detect dynamic cyber-threats, anomalies, and malicious activity within the network. However, dimension reduction becomes a more challenging task during machine learning model training as the amount of data increases. To address the issue, proposing the EFSMDCD method. This method selects optimal feature set in very short time and reduce the computational burden on machine. Moreover, our model evaluated on publicly available CICIDS2017 dataset. And our method achieved 99.95% accuracy with random forest algorithm using optimal feature subset. The application of Random Forest for DDoS detection has shown promising results, with high accuracy and robustness against complex attack patterns. The outcomes of this investigation offer a solid foundation for developing advanced DDoS detection systems using machine learning.