异常的网络环境往往可以通过观察网络流量的变化而提前发现,因此,如何实时准确地判定流量异常便成为网络异常检测中的重点。本文将自适应滤波理论引入网络流量异常检测,提出一种基于自适应AR模型的网络流量异常检测方法,该方法收敛快、精度高,能降低误报率,从而使异常检测更准确。 Abnormal network environment can often be found by observing the changes of network traffic in advance. Therefore, how to judge the abnormal traffic in real time becomes the focus of network anomaly detection. In this paper, adaptive filtering theory is introduced into the network traffic anomaly detection. An adaptive AR model based network traffic anomaly detection method is proposed. The proposed method has the advantages of fast convergence, high precision and low false alarm rate, which can make the anomaly detection more accurate.