摘要
针对当前网络安全态势要素提取方法未考虑多特征降维态势信息,导致网络安全态势要素分类正确率和召回率较低,误警率较高的问题,提出了基于多特征降维的网络安全态势要素提取方法。搭建网络安全态势要素提取架构,获取网络安全态势信息,采用非负矩阵分解算法,多特征提取与降维态势信息,构造并训练超球体分类器,获得态势要素信息分类函数,制定网络安全态势要素提取程序,实现网络安全态势要素提取。实验结果表明,在不同数据分布情况下,提出方法的网络安全态势要素分类正确率和召回率较高,能够有效降低误警率。
The neglect of multi-feature dimensionality reduction situation information leads to the low classification accuracy, recall rate and high false alarm rate of traditional network security situation elements extraction methods. This paper reports a network security situation element extraction method based on multi-feature dimensionality reduction. Firstly, the network security situation element extraction architecture was founded for obtaining the network security situation information. Secondly, the nonnegative matrix decomposition algorithm was introduced to extract multiple features and reduce the dimension of situation information. Then, the hypersphere classifier was constructed and trained to obtain the situation element information classification function. Eventually, the network security situation element extraction program was formulated, and the network security situation element extraction was achieved. The experimental results show that this method has high classification accuracy and recall rate of network security situation elements and a low false alarm rate.
作者
陈明
宋洋
CHEN Ming;SONG Yang(Changchun University of Technology,Changchun Jilin 130012,China;College Of Optical And Electronical Information Changchun University of Science and Technology,Changchun Jilin 130114,China)
出处
《计算机仿真》
北大核心
2022年第1期339-342,347,共5页
Computer Simulation
关键词
多特征降维
网络安全
安全态势
要素提取
Multi-feature dimension reduction
Network security
Security posture
Element extraction
