Zero-Day Attack Detection using Ensemble Technique

The zero-day attacks exploit the new vulnerabilities in the system or old vulnerabilities in a new way. Zero-day attacks are sustainable in the system exploiting the system until detected or until the patch is released, this creates a dire need to detect zero-day attacks in the system. The model in this proposed paper is an outlier-based model trained using benign and known attack traffic to detect traffic of unknown attacks. The proposed system successfully detected most of the unknown attack traffic, achieving higher results in Web, Infiltration, SSH, FTP, and Botnet attacks.