基于残噪预测的网络流量异常检测算法

A Detection Method For Network Traffic Anomalies Based On Rasidual Prediction

提出一种新的网络流量异常检测算法,拟对观察值序列进行AR模型的拟合,并以残噪函数值作为统计量,用指数平滑方法来实现前一个统计量对后一个统计量的预测,从而完成对异常流量的判断。该算法与GLR检测算法相比更为及时高效,比直接用指数平滑技术对观测值进行检测更为准确可靠,在OPNET模拟的网络流量异常检测中,该算法性能较为理想。

This paper proposes a new algorithms which is based on combining the linaear model and the method about smooth exponential. This algorithm is more effitiently comparing with GLB method and more reliable comparing with smooth exponential method. This algorithm has been proved is effectively in the detection for network traffic anomalies which is simulated with OPNET.