| 引用本文： |
-
肖禹名,赵小林,刘振岩,宋策,常悦.基于李群的网络系统行为风险计算方法[J].信息安全学报,已采用 [点击复制]
- Xiao Yuming,Zhao Xiaolin,Liu Zhenyan,Song Ce,Chang Yue.A Method for Calculating Behavioral Risk in Network Systems Based on Lie Group[J].Journal of Cyber Security,Accept [点击复制]
|
|
| 摘要: |
| 随着互联网在生活中的广泛应用，越来越多的软件开始收集更多用户信息以改进用户体验，而这些存放在服务器上的用户隐私，同时也存在着极大的泄露风险。对网络风险进行实时评估，不仅可以关注到网络状态的变化，也有利于随时调整网络防御手段，及时防御网络攻击，减小攻击损失。传统上，网络风险评估往往采用统计计算的方法来进行，本文通过采用数学中的李群模型，对网络系统进行数学建模，从而提出了一种实时计算网络风险的新算法。本文用李群运动学描述网络中的攻击行为。通过将网络系统中由指标和拓扑组成的矩阵映射到李群，给出攻击行为路径以及网络攻防的数值定义，使用测地线计算李群中元素的距离，作为网络风险指标，并提出了相应的网络风险损害评估方法，从而将网络风险量化，实现对网络安全状态的实时评价。为了检验这种网络安全风险评估的有效性，本文使用现有数据集，并编写代码进行了相关实验，以评估这种方法的适用性与效率。实验结果证实，该基于李群的网络系统行为风险计算方法对于网络攻防风险值的客观量化计算是有效的，可以实现对网络风险的定量评估。在与其他的机器学习算法相比时，各指标上均没有明显差距，而其具有的一些特点则具备被深度开发的潜质。 |
| 关键词: 网络安全 李群 数据安全 风险计算 行为风险 |
| DOI：10.19363/J.cnki.cn10-1380/tn.2024.12.01 |
| 投稿时间：2023-11-30修订日期：2024-02-01 |
| 基金项目:国家重点研发课题《健身知识和线上指导数据安全与隐私保护技术研究》（课题编号2022YFC3600404） |
|
| A Method for Calculating Behavioral Risk in Network Systems Based on Lie Group |
|
Xiao Yuming, Zhao Xiaolin, Liu Zhenyan, Song Ce, Chang Yue
|
| (Beijing Institute of Technology) |
| Abstract: |
| With the widespread use of the Internet in our lives, more and more software is beginning to collect more user information to improve the user experience. However, there is also a significant risk of leakage of user privacy stored on servers. Real-time assessment of network risks can not only help to monitor changes in network status, but also facilitate the adjustment of network defense measures at any time, and timely defense against network attacks, reducing attack losses. Traditionally, network risk assessment often uses statistical methods to calculate. This article proposes a new algorithm for real-time calculation of network risks by using the Lie group model in mathematics to mathematically model network systems. This article uses Lie group kinematics to describe attack behaviors in the network. By mapping the matrix composed of indicators and topology in the network system to the Lie group, it gives the numerical definition of attack behavior paths and network attack and defense. Using geodesic to calculate the distance between elements in the Lie group, it serves as a network risk indicator, and proposes a corresponding network risk damage assessment method, thus quantifying network risks and achieving real-time evaluation of network security status. In order to test the effectiveness of this network security risk assessment method, this paper uses existing datasets and writes code to conduct relevant experiments to evaluate the applicability and efficiency of this method. The experimental results confirm that the Lie group-based network system behavior risk calculation method is effective for the objective quantitative calculation of network attack and defense risk values, and can achieve quantitative assessment of network risks. Compared with other machine learning algorithms, there is no significant difference in various indicators, and some of its characteristics have the potential for further development. |
| Key words: network security, Lie Group,data security , risk calculation, behavioral risk |
