A Survey on Anomaly Detection in Network with ML Techniques

Anomaly detection is the process of finding malicious data that may occur in computer networks. There are multiple anomaly detection techniques available in practice. As the new technologies are growing fast in networking because of the connecting devices. A novel approach can be introduced using machine learning, which may be flexible enough to detect intrusions in a network structure. Nowadays, the dataset is imbalanced, so the anomaly detection model results are average. In this paper, anomaly detection is summarised based on different machine learning algorithms. We also reviewed the implementation and challenges faced in the ML technique to detect anomalies like DoS and DDoS attacks. The performance is reviewed from the survey only through the ML metrics. Due to the high range of multimedia data, it is essential to concentrate on Software Defined Networks rather than traditional networks. The advantages and disadvantages of various ML approaches to detecting all kinds of general anomalies are discussed in this paper.