A Lightweight Online Network Anomaly Detection Scheme Based on Data Mining Methods

This paper presents our preliminary work in network anomaly detection. The experimental results demonstrate an inspiring and promising trend for lightweight on-line network anomaly detection, which is rather meaningful for the ever-increasing network traffic and the accompanied network threats. In our future work, we will further verify and optimize our methods in terms of the concrete applications, as well as deploying it in our national backbone network to detect anomalies such as DoS, DDoS, probe, spam, etc.