基于决策树的SDN网络入侵分类检测模型  

SDN network intrusion classification detection model based on decision tree

作　　者：李道全 杨乾乾 鲁晓夫 LI Dao-quan;YANG Qian-qian;LU Xiao-fu(School of Information and Control Engineering,Qingdao University of Technology,Qingdao 266520,China)

机构地区：[1]青岛理工大学信息与控制工程学院,山东青岛266520

出　　处：《计算机工程与设计》2022年第8期2146-2152,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61572269)。

摘　　要：软件定义网络(SDN)由于只有一个控制点,更容易受到网络攻击,为更好检测和防范网络攻击,提出一种基于决策树的SDN网络入侵分类检测模型。通过引入类间中心距离,分别计算各类和其它类别的类间中心距离,以此作为判断各个类别分离程度的依据,确定类别的先后分离顺序。通过将多分类任务分解为多个两类分类问题,构造多分类决策树模型。实验结果表明,基于类间中心距离的SDN网络入侵分类检测算法具有良好的检测性能。Software-defined network(SDN)has only one control point and is more vulnerable to network attacks.To detect and prevent network attacks,a SDN network intrusion classification detection model was proposed based on a decision tree.By introducing the distance between classes,the class center distances of each class and other classes were calculated separately,which was used as the basis for judging the degree of separation of various classes and determining the separation order of the classes.By decomposing the multi-classification task into multiple two-class classification problems,a multi-class decision tree model was constructed.Experimental results also reflect the superiority of the SDN network intrusion classification detection algorithm based on the distance between classes.

关 键 词：软件定义网络 决策树 机器学习 入侵分类检测 类间距离 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]