系统仿真学报 ›› 2021, Vol. 33 ›› Issue (6): 1288-1296.doi: 10.16182/j.issn1004731x.joss.20-0112

• 仿真建模理论与方法 • 上一篇    下一篇

结合栈式自编码及长短时记忆的入侵检测研究

林硕1, 安磊1, 高治军1,*, 单丹1, 尚文利2,3,4   

  1. 1.沈阳建筑大学 信息与控制工程学院,辽宁 沈阳 110168;
    2.中国科学院沈阳自动化研究所 工业控制网络与系统研究室,辽宁 沈阳 110168;
    3.中国科学院网络化控制系统重点实验室,辽宁 沈阳 110016;
    4.中国科学院机器人与智能制造创新研究院,辽宁 沈阳 110016
  • 收稿日期:2020-03-09 修回日期:2020-05-19 出版日期:2021-06-18 发布日期:2021-06-23
  • 通讯作者: 高治军(1978-),男,博士,副教授,研究方向为智慧建筑、网络安全。E-mail:gzj@sjzu.edu.cn
  • 作者简介:林硕(1981-),男,博士,副教授,研究方向为排产优化、信息安全。E-mail:farewell_lin@163.com
  • 基金资助:
    国家自然科学基金(61773368); 辽宁省教育厅科学技术项目(Injc201912); 辽宁省教育厅青年科技人才“育苗”项目(Inqn201912)

Research on Intrusion Detection Based on Stacked Autoencoder and Long-short Memory

Lin Shuo1, An Lei1, Gao Zhijun1,*, Shan Dan1, Shang Wenli2,3,4   

  1. 1. Faculty of Information and Control Engineering, Shenyang Jianzhu University, Shenyang 110168, China;
    2. Department of Industrial Control Network and System, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110168, China;
    3. Key Laboratory of Network Control System, Chinese Academy of Sciences, Shenyang 110016, China;
    4. Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China
  • Received:2020-03-09 Revised:2020-05-19 Online:2021-06-18 Published:2021-06-23

摘要: 针对网络攻击越来越隐蔽,且具有智能化和复杂化的特点,浅层的机器学习已经无法及时应对,提出了一种基于SDAE(Stacked Denoising Autoencoder)和LSTM(Long Short-Term Memory)相结合的深度学习方法。通过堆叠深层的SDAE智能逐层抽取网络数据的分布规则,结合各个编码层的系数惩罚和重构误差对高维数据进行多样性异常特征提取。结合LSTM的记忆功能和强大的序列数据学习能力进行学习分类。在UNSW-NB15数据集上进行了实验,通过调整时间步长进行分析,实验结果表明,该模型具有检测准确率高、误报率低的优点。

关键词: 深度学习, 入侵检测技术, 栈式降噪自编码器, 长短时记忆网络

Abstract: As network attacks increasingly hidden, intelligent and complex. Simple machine learning cannot deal with attacks timely. A deep learning method based on the combination of SDAE and LSTM is proposed. Firstly, the distribution rules of network data are extracted intelligently layer by layer by SDAE, and the diverse anomaly features of high-dimensional data ate extracted by using coefficient penalty and reconstruction error of each coding layer. Then, LSTM’ s memory function and the powerful learning ability of sequence data are used to classify learning depth. Finally, the experiments are carried out with the UNSW-NB15 data set, which is analyzed by adjusting the time step. The experimental results show that the model has higher detection accuracy and lower false alarm rate.

Key words: deep learning, Intrusion Detection System(IDS), Stacked Denoising Autoencoder (SDAE), Long Short-Term Memory (LSTM)

中图分类号: