Analysis of AI and Optimization Approaches for Cyber System Anomaly Identification

With the surge in cyber system deployments, anomaly detection has become imperative to counter rising security threats. Traditional rule-based and statistical models often falter against dynamic attack patterns, motivating the need for adaptive and scalable detection approaches. This paper presents a comprehensive survey on advanced anomaly detection techniques under the proposed framework Adaptive Detection of Cyber Anomalies using ML/DL and Optimization Strategies (ADCAMOS). Primary objective is to evaluate the efficacy of Machine Learning (ML), Deep Learning (DL), and nature-inspired optimization methods in identifying sophisticated cyber anomalies. This study reviews key ML/DL models, feature selection mechanisms, and the role of optimization algorithms in fine-tuning detection performance. Experimental insights demonstrate how these methods improve anomaly detection accuracy and robustness in real-time environments. Furthermore, the impact of adversarial attacks on model resilience is analyzed, highlighting strategies to fortify cyber system reliability and security.