摘要
通过对网络病毒的动态交互约束抑制设计,实现对病毒入侵特征的有效识别。传统方法采用模糊网络入侵状态特征向量分解方法实现病毒约束抑制,当病毒入侵为一种非平稳随机信息矢量时,对其识别性能不好。提出一种基于互信息特征提取的网络病毒动态交互约束算法。构建网络病毒入侵的信号分析模型,并进行数据采集,采用数模转换方法进行病毒数据离散采样转换,采用重采样和机器学习结合方法,进行了链路漏洞检测,填补了Web防火墙的漏洞,采用三次B样条小波进行互信息特征提取的结果是渐近最优的,利用互信息特征作为检测系统的输入,进行病毒数据提纯处理,基于平均互信息特征提取算法实现特征建模和提取,实现病毒动态交互约束。仿真结果表明,该算法能使得病毒数据在时频空间上得到较明显的聚焦,频谱峰值突出,提高了病毒特征有效识别率。
The effective recognition of the characteristics of the virus invasion is realized by restraining thedynamic interaction of the network virus. In the traditional method, the method of fuzzy network intrusiondetection is used to implement the virus constraint. When the virus is invaded as a kind of nonstationaryrandom information vector, the identification performance is not good. A dynamic interactiveconstraint algorithm for network virus based on mutual information feature extraction is proposed. Thesignal analysis model of network virus invasion is constructed, and the data acquisition is carried out byusing digital mode conversion method. The data are analyzed by using the method of heavy sampling andmachine learning. The results of Web firewall are extracted. Using three B spline wavelet, the mutualinformation is used as the input. The simulation results show that the algorithm can make the virus datain time frequency space to get a more obvious focus, and the spectrum peak value is outstanding, whichcan improve the recognition rate of the virus.
出处
《科技通报》
北大核心
2016年第5期91-95,共5页
Bulletin of Science and Technology
基金
河南省自然科学基金研究项目资助(152300410145)
河南省教育厅科学技术重点研究项目(15B520012)
关键词
病毒
网络安全
动态交互
互信息特征
virus
network security
dynamic interaction
mutual information feature
