AI-Driven Network Anomaly Detection for Enhanced Cybersecurity and Performance

Threats of the network that have ever/never happened could be anomalies. Protecting networks against malicious access has always been challenging and despite a long time study into this matter, it never becomes easier. Because of the evolutionary step of the network through new technologies as well as fast growth of connected devices, attacks on the network are getting versatile too. Compared to conventional intrusion detection methods, machine learning is a new flexible approach to detect intrusions in the network and applicable to any kind of network structure. This paper discusses the challenges of anomaly detection in both the traditional network and the next generation network. The implementation of machine learning in anomaly detection under various network contexts. The process of every machine learning class is defined and the methodologies along with the advantages. A summary of the comparison of which uses different models of machine learning in the process is also presented.