中文摘要：

基于Hurst指数进行异常检测的模型多采用固定阈值的方法,不能很好的适应动态变化的网络环境。针对该问题设计了一种基于动态阈值的检测方法,该方法在采用Hurst指数分析的基础上,通过EWMA和滑动窗口模型控制有效数据的个数并根据网络的变化动态调整检测阈值,提高了模型的检测能力。实验结果表明,在采用动态阈值进行DDOS异常检测时具有较高的检测率。

英文摘要：

Most anomaly detection model based on Hurst index use fixed threshold,but it doesn＇t well adapt to the dynamic changing network.To sdve the problem,a dynamic threshold detection method is proposed.After using wavelet analysis in the Hurst value,according to changes in the network,it controls the number of valid data to dynamically adjust the threshold value based on sliding window and EWMA.Thus,the detection ability of the model is improved.Experimental results show that the method of dynamic thresholds have a higher detection rate in DDOS detection.