摘要
针对网络攻击类型多,高维网络数据提取困难及入侵检测算法需具备较高自调节能力的问题,从能量变换寻优与集成算法的角度,提出了一种基于自调节深度信念网络的入侵检测算法,并采用国际通用入侵检测数据集进行检测效果验证.结果表明,与当前主流的入侵检测算法相比,该算法能够提高入侵检测的准确率,具有全局性与稳定性.
Whereas there exist varied kinds of cyber attacks, difficulty of extraction of high-dimensional network data and need for higher self-regulation capabilities of intrusion detection algorithms, the writer of this paper argues for an intrusion detection algorithm, a self-adjusting deep belief network(SADBN), from the optimization of energy transformation and integration of algorithms. Attesting via NSLKDD shows that this algorithm can improve the accuracy of intrusion detection.
作者
逯玉婧
LU Yu-jing(Department of Personnel,Shijiazhuang University of Applied Technology,Shijiazhuang,Hebei 050081,China)
出处
《石家庄职业技术学院学报》
2021年第2期1-4,共4页
Journal of Shijiazhuang College of Applied Technology
基金
石家庄职业技术学院院级科研基金一般课题(20YQ1009)。
关键词
自调节
深度信念网络
入侵检测技术
受限玻尔兹曼机
特征提取
self-regulation
deep belief networks
intrusion detection technology
restricted Boltzmann machine
extraction