Abstract—Detecting a variety of anomalies in computer network, especially zero-day attacks, is one of the real challenges for both network operators and researchers. An efficient technique detecting anomalies in real time would enable network operators and administrators to expeditiously prevent serious consequences caused by such anomalies. We propose an alternative technique, which based on a combination of time series and feature spaces, for using machine learning algorithms to automatically detect anomalies in real time. Our experimental results show that the proposed technique can work well for a real network environment, and it is a feasible technique with flexible capabilities to be applied for real-time anomaly detection.
Index Terms—Multivariate normal distribution, nearest neighbor, one-class support vector machine, unsupervised learning.
Kriangkrai Limthong is with the Department of Computer Engineering, School of Engineering, Bangkok University, Pathumtani 12120, Thailand (e-mail: kriangkrai.l@bu.ac.th). He is also now with the Department of Informatics, Graduate University of Advanced Studies (Sokendai), Chiyoda-ku, Tokyo 101-8430, Japan (e-mail: krngkr@nii.ac.jp).
[PDF]
Cite:Kriangkrai Limthong, "Real-Time Computer Network Anomaly Detection Using Machine Learning Techniques," Journal of Advances in Computer Networks vol. 1, no. 1, pp. 1-5, 2013.
