摘要
针对传统入侵检测方法在检测效果上存在检测率低、误报率高等问题,将深度信念网络应用于入侵检测中,提出一种基于灰狼算法(GWO)的深度信念网络(DBN)入侵检测方法。对NSL-KDD进行预处理提高其鲁棒性,搭建深度信念网络检测模型,引入灰狼算法对其隐含层节点数进行全局寻优,近似得到DBN最佳网络结构,利用所得数据集进行验证分析。实验结果表明,与几种常用的入侵检测算法相比,经过灰狼算法优化后的深度信念网络提高了入侵检测的检测率,降低了误报率,为入侵检测提供了新的依据。
Aiming at the low detection rate and high false positive rate of traditional intrusion detection methods,the deep belief network was applied to intrusion detection and a deep belief network(DBN)intrusion detection method based on grey wolf algorithm(GWO)was proposed.The NSL-KDD data was preprocessed to improve its robustness,the DBN training model was constructed and the gray wolf algorithm was introduced to globally optimize the DBN hidden layer nodes and to approximate the DBN optimal network structure.The resulting data set was used for verification analysis.Experimental results show that compared with several commonly used intrusion detection algorithms,the deep belief network optimized using the grey wolf algorithm improves the detection rate of intrusion detection and reduces the false positive rate,which provides a new basis for intrusion detection.
作者
夏景明
丁春健
谈玲
XIA Jing-ming;DING Chun-jian;TAN Ling(College of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China;College of Computer and Software,Nanjing University of Information Science and Technology,Nanjing 210044,China)
出处
《计算机工程与设计》
北大核心
2020年第6期1534-1539,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(41505017)
江苏省自然科学基金项目(BK20160951)。
关键词
入侵检测
灰狼算法
深度信念网络
全局寻优
网络结构
intrusion detection
grey wolf algorithm
deep belief network
global optimization
network structure
