网络安全态势的评估与预测技术研究

发布时间：2018-12-14 16:17

【摘要】：随着信息时代的到来,互联网得以快速发展和普及。但是与此同时,也给人们的生活带来了严重的危害,那便是网络安全事件频繁发生。尽管现在有各种各样的网络安全设备用来保护互联网安全,但是由于他们是针对不同的安全问题而设计的,具有专一性和不同的侧重点,所以这就导致他们不能很好地评估和预测整个网络的安全状态。在本论文中,通过对当前网络安全态势评估和预测方法的详细分析和研究,表明当前如何提高网络安全态势预测的预测精度和提高预测的收敛速度仍然是需要解决的热点问题。本文针对提高网络安全态势预测的预测精度问题,提出了基于二分K-均值的径向基神经网络安全态势预测模型,该方法使用二分K-均值聚类算法确定径向基神经网络的数据中心和扩展函数,弥补了径向基神经网络数据中心难以确定的缺陷。经实验验证,该方法在一定条件下提高了预测精度。本文针对提高网络安全态势预测的收敛速度问题,提出了基于改进的人工免疫的网络安全态势预测方法,该方法使用了等价划分区间的策略来改进人工免疫算法,弥补了在随机生成初始抗体的过程中产生的数据冗余性缺陷。经实验验证,该方法提高了预测的收敛速度。
[Abstract]:With the arrival of the information age, the Internet has been rapidly developed and popularized. But at the same time, it also brings serious harm to people's life, that is, network security incidents occur frequently. Although there are a variety of network security devices available to protect Internet security, because they are designed for different security issues, they are specific and have different emphases. So this leads to their inability to evaluate and predict the security of the entire network. In this paper, through the detailed analysis and research on the current network security situation assessment and prediction methods, it is shown that how to improve the prediction accuracy and convergence rate of network security situation prediction is still a hot issue to be solved. In order to improve the prediction accuracy of network security situation prediction, a radial basis function neural network security situation prediction model based on dichotomous K-means is proposed in this paper. In this method, the binary K-means clustering algorithm is used to determine the data center and the expansion function of the radial basis function neural network, which makes up for the difficulty of determining the data center of the radial basis function neural network. The experimental results show that this method can improve the prediction accuracy under certain conditions. In order to improve the convergence rate of network security situation prediction, a network security situation prediction method based on improved artificial immune is proposed in this paper. It makes up for the data redundancy in the process of generating the initial antibody randomly. Experimental results show that the proposed method improves the convergence rate of prediction.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.08

【参考文献】