摘要
针对现有网络安全技术不能准确地对网络未来安全态势进行预测的问题,提出一种基于回声状态网络(ESN)进行态势预测的方法。对数据量庞大的安全态势值,利用回声状态网络可有效处理非线性系统辨识以及混沌时间序列预测问题。实验表明,由于训练简单,可根据不同网络环境自动优化网络参数,且算法本身在混沌时间序列预测方面的先天优势,使得ESN算法比传统神经网络方法在网络态势预测准确率方面有明显改善。
Since the existing network security technology can’t accurately solve the problem of net-work future security situation forecast,a trend prediction method is presented based on ESN echo state network.For a large number of security situational values,using ESN （echo state network）nonlinear system identification and the chaotic time series prediction problems can be effectively deal with.The ex-perimental results show that this method can automatically optimize the network parameters according to different network environment because of its simple training.And this algorithm itself has a congenital advantage in chaotic time series prediction,thus it has an obvious improvement in network trend predic-tion accuracy.
出处
《滨州学院学报》
2015年第2期81-85,共5页
Journal of Binzhou University
基金
国家自然科学基金资助项目(61202390)
关键词
回声状态网络
网络安全态势感知
态势值预测
ESN
ESN（echo state networks）
network security situation awareness
situational value pre-diction
