摘要
现有校园网络入侵数据自主防御方法通常需要依赖足够数量样本的分析,但由于网络攻击技术的提升,大量新型攻击技术涌现,初始训练样本数量有限严重影响了校园网络入侵数据的检测性能和自主防御性能,提出了基于数据挖掘的网络入侵数据自主防御方法。方法将采集获得的校园网络数据进行离散连续化、标准化和归一化等预处理;通过采用数据挖掘方法中的模糊C均值聚类随机选取一个聚类中心,迭代目标函数,寻找目标函数的最小值,并不断调整聚类中心和隶属度,获得校样本最佳类别,完成校园网络数据集聚类;在此基础上通过度量聚类后各个数据集簇的异常度判断是否为入侵数据;基于校园网络入侵数据检测结果构建了一个三维立体自主防御架构,实现了校园网络入侵数据自主防御。仿真结果表明,所提方法能够克服了当前方法的弊端,实现了校园网络入侵数据的准确检测和完全自主防御。
The existing self-defense methods of campus network intrusion data usually rely on a sufficient number of samples. However, the limited number of initial training samples seriously influences the detection performance and autonomous defense of campus network intrusion data. In this article, a method of autonomous defense for network intrusion data based on data mining was proposed. This method performed the pretreatment of discrete continuation, standardization and normalization on the collected campus network data. Our method used the fuzzy C-means clustering in data mining method to randomly select a clustering center. By iterating the objective functions, we could find the minimum value of objective functions. Continuously, we adjusted the clustering center and membership degree to get the best category of samples, and thus completing the campus network dataset clustering. On this basis, we measured the abnormality of each data cluster, so as judge whether it was intrusion data. Based on the result of campus network intrusion data detection, we built 3 D autonomous defense architecture. Thus, we realized the autonomous defense of campus network intrusion data. Simulation results show that the proposed method can overcome the shortcomings of current method and achieve the accurate detection and complete autonomous defense of campus network intrusion data.
作者
张代华
沈勇
章翔飞
王兵
ZHANG Dai-hua;SHEN Yong;ZHANG Xiang-fei;WANG Bin(School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212003,China)
出处
《计算机仿真》
北大核心
2020年第10期263-267,共5页
Computer Simulation
基金
2018年江苏省教育信息化研究课题(20180013)。
关键词
数据挖掘
网络入侵
数据
自主防御
Data mining
Network intrusion
Data
Autonomous defense
