Anomaly-based NIDS: A review of machine learning methods on malware detection

Abstract

The increasing amount of network traffic threat may originates from various sources, that can led to a higher probability for an organization to be exposed to intruder. Security mechanism such as Intrusion Detection System (IDS) is significant to alleviate such issue. Despite the ability of IDS to detect, some of the anomaly traffic may not be effectively detected. As such, it is vital the IDS algorithm to be reliable and can provide high detection accuracy, reducing as much as possible threats from the network. Nonetheless, every security mechanism has its weaknesses that can be exploited by intruders. Many research works exists, that attempts to address the issue using various methods. This paper discusses a hybrid approach to network IDS, which can minimize the malicious traffic in the network by using machine learning. The paper also provides a review of the available methods to further improve Anomaly-based Network Intrusion Detection System.