RECENT ADVANCES IN ALGORITHMS FOR DETECTING AND MITIGATING NETWORK TRAFFIC ANOMALIES

Network traffic anomalies pose significant threats to the security, performance, and reliability of information and communication systems (ICS). As networks become more complex and generate massive volumes of real-time data, detecting and mitigating these anomalies has become increasingly challenging. This paper explores recent advancements in algorithms for detecting and mitigating network traffic anomalies. Key approaches discussed include machine learning-based methods (deep learning, ensemble learning, federated learning), statistical techniques (RPCA, wavelet analysis), graph-based methods (GNNs, community detection), and anomaly mitigation strategies leveraging Software-Defined Networking (SDN) and reinforcement learning. The study highlights the strengths and limitations of each approach, along with their practical use cases, and addresses current challenges such as high false positives, scalability, and the need for real-time detection.