中文摘要：

针对网络流量发生异常时产生的突变特征,提出了一种基于突变级数的网络流量的异常检测方法.该方法首先计算网络流量的特征量,选择其中能显著性反映网络流量自相似性、非线性、非平稳性及复杂的动力学结构特性的特征量;然后将其作为突变理论的控制变量,利用蝴蝶突变模型的突变级数对网络流量异常进行检测.实验结果表明该方法具有较高的检测率和较低的误检率.

英文摘要：

Aimed at the catastrope characteristic when there are anomalies of network traffic happened,a network traffic anomaly detection approach based on the catastrophe progression theory was proposed.Some features of network traffic were calculated.The features significantly reflecting the self-similarity,non-linear,non-stationary nature and complexity of the dynamic structure of network traffic were chosen as the control variables of catastrophe theory.Then we used the catastrophe progression corresponding to the butterfly catastrophe model to detect the anomaly of network traffic.The experimental results show that the proposed approach has a low false alarm ratio and a high detection accuracy.