摘要
在相关分析原理基础上,建立网络观测系统的模型,确定了网络流量的相关性特征,然后将多尺度相关分析方法引入系统检测中,提出了一种流量异常检测与定位方法。实验表明,与残差相关分析方法相比较,克服了干扰过多或者丢失信号细节成分两种不同性质的错误,该方法适合于突变信号的处理。
The paper presents the detection and location algorithm of abnormal network Traffic based on the principle of correlation analysis. Firstly, the network observation system model is constructed, and the correlation characteristic of the network flow is determined. Secondly, the multi-scale analysis is defined for the system detection, overcomes two different kinds of property errors which resulted the excessive signal interference or detail components compared with the residual correlation analysis method. Finally, expression shown the algorithm is suitable for the mutation signal processing.
出处
《齐齐哈尔大学学报（自然科学版）》
2014年第4期15-17,43,共4页
Journal of Qiqihar University(Natural Science Edition)
基金
齐齐哈尔市科技局科技公关项目(GYGG-201213-1)
关键词
多尺度相关分析
异常网络流量
检测与定位
multi-scale correlation analysis
abnormal network traffic
detection and location
