摘要
对多组件网络节点漏洞进行准确识别,有利于节点漏洞的高效检测修补,提高多组件网络运行安全性。对多组件网络节点漏洞的准确识别,需要得到多组件网络节点漏洞属性特征集合,构建漏洞识别适应度函数,完成节点漏洞的准确识别。传统方法选取一组最好参数作为神经网络权值,构建网络节点漏洞识别模型,但忽略了对漏洞识别的适应度函数的求取,导致识别精度偏低。提出基于HIWO与SVM的多组件网络节点漏洞识别方法,通过知识约简提取网络节点漏洞属性特征,对连续的网络节点漏洞数值属性特征离散化过程进行控制,得到具有代表性的多组件网络节点漏洞属性特征集合;将节点漏洞检测正确率当作HIWO种子适应度函数,找到SVM最优参数值,构建网络节点漏洞识别模型。仿真表明,上述方法对节点漏洞特征向量的提取效果较为理想,可准确识别多组件网络节点漏洞,具有可实践性。
This article proposes a method for identifying node vulnerability of multi-component network based on HIWO and SVM. Based on knowledge reduction, this method extracted the attribute characteristic of network node vulnerability and controlled the consecutive discretization process of numerical attribute characteristic of network node vulnerability to get the representative set of multi-component network node vulnerability attribute characteristic. Then, the detection accuracy of node vulnerability was regarded as HIWO fitness function. Thus, the optimal parame- ters of SVM were found and the identification model of network node vulnerability was built. Simulation results show that the proposed method is effective for extracting feature vector of node vulnerability. Meanwhile, this method can accurately identify multi-component network node vulnerability, which is practicable.
作者
黄熙岱
HUANG Xi-dai(Faculty of Mathematics and Computer Science,Guangdong Ocean University,Guangdong Zhanjiang 524088,Chin)
出处
《计算机仿真》
北大核心
2018年第7期323-326,共4页
Computer Simulation
关键词
多组件
网络节点
漏洞识别
Multiple components
Network nodes
Vulnerability identification
