摘要
考虑到软件定义网络异常流量分类受到网络复杂特性的影响,导致分类效果变差,提出了基于K-means聚类的软件定义网络异常流量分类研究。利用权重属性,划分了软件定义网络异常流量的频度,在网络异常流量的分布形式转化过程基础上,根据软件定义网络异常节点出现的概率,计算了异常流量的特征值,提取出软件定义网络异常流量特征,根据软件定义网络异构值差的度量,初始化软件定义网络的参考向量值,通过更新分类属性的邻域半径,计算网络异常流量的参考向量,选择出软件定义网络异常流量分类属性,利用K-means聚类算法过滤软件定义网络异常流量,对软件定义网络异常流量进行检索,通过定义网络异常流量分类的目标函数,利用K-means聚类算法理论,构建软件定义网络异常流量的加权临界函数,结合对角矩阵的求解,设计了软件定义网络异常流量分类原理,实现了软件定义网络异常流量的分类。实验结果表明,文中分类方法的查全率、差准率较高,适应度以及收敛性能较好。
Considering that the classification of software defined network abnormal traffic is affected by the complexity of the network,resulting in the poor classification effect,a research on software defined network abnormal traffic classification based on K-means clustering is proposed.Using the weight attribute,the frequency of software defined network abnormal traffic is divided.Based on the transformation process of the distribution form of network abnormal traffic,the eigenvalues of abnormal traffic are calculated according to the probability of network abnormal nodes defined by software,the characteristics of software defined network abnormal traffic are extracted,and the measurement of network heterogeneous value difference is defined according to software.Initialize the reference vector value of the software defined network,calculate the reference vector of network abnormal traffic by updating the neighborhood radius of the classification attribute,select the software defined network abnormal traffic classification attribute,filter the software defined network abnormal traffic by using K-means clustering algorithm,retrieve the software defined network abnormal traffic,and define the objective function of network abnormal traffic classification.Using the theory of K-means clustering algorithm,the weighted critical function of software defined network abnormal traffic is constructed.Combined with the solution of diagonal matrix,the classification principle of software defined network abnormal traffic is designed,and the classification of software defined network abnormal traffic is realized.The experimental results show that the classification method in this paper has better classification effect in recall rate,difference rate and fitness index.
作者
王彬彬
WANG Bin-bin(Fuyang Preschool Teachers College,Anhui Fuyang 236015,China)
出处
《齐齐哈尔大学学报（自然科学版）》
2022年第2期50-55,90,共7页
Journal of Qiqihar University(Natural Science Edition)
基金
阜阳幼儿师范高等专科学校校级质量工程——常态化疫情防控下的高职计算机基础教学探讨(ZLGC2020JY011)。
关键词
K-MEANS聚类
软件定义网络
异常流量
分类方法
特征提取
属性选择
K-means clustering
software defined network
abnormal flow
classification method
feature extraction
attribute selection
