Comparative Learning Approaches for Zero Day Attack Anomaly Detection Systems

¹ Sochima Godson Alli-Okoro
² Chukwudi Igbe
³ Eleberi Ebele Leticia

^1,2,3 Department of Computer Science, Imo State University, Owerri.

Abstract
This opinion research topic explored Comparative Learning Approaches for Zero-Day Anomaly Detection Systems, a critical component of modern cyber-security. The escalating sophistication of cyber threats has rendered traditional security measures inadequate in detecting and mitigating zero-day attacks (ZDAs), necessitating this research. The purpose of this study is to evaluate learning approaches for anomaly detection in ZDAs and provide insights for improving cyber-security posture. The research conceptualized ZDAs in modern cyber-security, highlighting their impact on organizations. The imperative of anomaly detection in ZDAs is discussed, emphasizing the need for effective detection mechanisms. Learning approaches for anomaly detection, including supervised, unsupervised, and hybrid methods, are critically evaluated. A comparative analysis of these approaches is conducted, revealing their strengths and limitations. Using qualitative method, the researchers discussed the issues raised in the research, revisiting the case for anomaly detection in ZDAs. The findings revealed that hybrid approaches, combining supervised and unsupervised learning, offer a promising solution for detecting ZDAs. Based on the findings, the researchers recommended, among others, that organizations prioritize the development of hybrid anomaly detection systems and invest in continuous learning and adaptation to emerging threats.

Key Words: Zero-Day Attacks, Anomaly Detection, Learning Approaches, Cybersecurity, Hybrid Approaches.

Comparative-Learning-Approaches-for-Zero-Day-Attack-Anomaly-Detection-Systems