SS-UDANN:面向高维复杂数据的时空特征分析方法
作者:
作者单位:

1.沈阳化工大学 a信息工程学院;2.沈阳工业大学 人工智能学院

中图分类号:

TP393???????

基金项目:

辽宁省自然科学基金项目(2023-MSLH-273);辽宁省科学技术计划项目(2023JH1/10400082);辽宁省人工智能创新发展计划项目(2023JH26/1030008);辽宁省科技创新平台建设计划项目([2022]36号)


SS-UDANN: Spatiotemporal Feature Analysis Method for High-Dimensional Complex Data
Author:
Affiliation:

1.School of Information Engineering, Shenyang University of Chemical Technology.;2.School of Artificial Intelligence, Shenyang University of Technology

  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献 [25]
  • | |
  • 引证文献
  • | |
  • 文章评论
    摘要:

    现有特征分析方法往往仅关注时间或空间单一特征,忽视特征间关联和分布差异,导致难以分析高维数据复杂结构的问题。针对上述问题本文提出一种面向高维复杂数据的时空特征分析方法(SCASRCAE-Unsupervised Domain-Adversarial Neural Network,SS-UDANN)。该方法使用稀疏卷积自编码器(sparse regularization convolutional auto-encoder,SRCAE)作为无监督特征提取器,应用于改进的领域对抗网络(Domain-Adversarial Neural Network,DANN),并引入轻量级跨纬度空间通道注意力(Spatial Channel Attention,SCA)机制,在低计算成本下提取时空相关特征;同时,为解决特征分布差异问题,采用最大均值差异(Maximum Mean Discrepancy,MMD)进行正则化,训练特征提取器与域鉴别器获取域不变特征,实现数据增强。以工业领域网络异常数据检测为验证背景,应用SS-UDANN分析CICIDS2018数据集的时空特征,精确率为98.66%。并对实验室油气集输全流程攻防靶场数据进行特征分析,精确率为95.39%,验证了SS-UDANN的有效性,表明本数据处理方法具有应用推广价值。

    Abstract:

    To address the limitations of existing methods that only focus on temporal or spatial features, this study proposes a novel approach for analyzing high-dimensional complex data using spatiotemporal features. The method, named Spatiotemporal Sparse Regularization Convolutional Autoencoder Unsupervised Domain-Adversarial Neural Network (SS-UDANN), employs a Sparse Regularization Convolutional Autoencoder (SRCAE) as an unsupervised feature extractor and incorporates an enhanced Domain-Adversarial Neural Network (DANN). A lightweight cross-dimensional Spatial Channel Attention (SCA) mechanism is integrated to extract spatiotemporal features efficiently while keeping computational costs low. Maximum Mean Discrepancy (MMD) is applied to regularize the feature extractor and domain discriminator, facilitating the extraction of domain-invariant features for effective data augmentation. Validation of SS-UDANN on the CICIDS2018 dataset for industrial network anomaly detection shows an accuracy of 98.66%. Additionally, the model achieves 95.39% accuracy when applied to data from a laboratory oil and gas full-process attack-defense testbed, further confirming the method"s effectiveness and demonstrating its potential for broader application in data processing.

    参考文献
    [1] Feng S, Zhao L, Shi H, et al. One-dimensional VGG Net for high-dimensional data[J]. Applied Soft Computing, 2023, 135: 110035.
    [2] Wang Z, Wang H, Wang S, et al. Linear mixed-effects model for longitudinal complex data with diversified characteristics[J]. Journal of Management Science and Engineering, 2020, 5(2): 105-124.
    [3] 冯凯, 董秀成, 刘栋博.基于多域融合与特征选择的手势识别研究[J].传感器与微系统,2023,42(05):37-40+44.Feng K,Dong C X, Liu D B. Research on gesture recognition based on multi-domain fusion and feature selection[J]. Transducer and Microsystem Technologies,2023,42(05):37-40+44.
    [4] 陶鹏,邓绍江.自适配权重特征融合的持续身份认证[J].重庆大学学报,2023,46(01):103-112.Tao P,Deng S J .Self-adaptive weight feature fusion for continuous authentication[J]. Journal of Chongqing University, 2023,46(01):103-112.
    [5] Zhang Y,Zhang E,Chen W. Deep neural network for halftone image classification based on sparse auto-encoder[J]. Engineering Applications of Artificial Intelligence,2016,50: 245-255.
    [6] Deng W, Li Y, Huang K, et al. LSTMED: an uneven dynamic process monitoring method based on LSTM and Autoencoder neural network[J]. Neural Networks, 2023, 158: 30-41.
    [7] Cheng P, Han M, Liu G. DESC-IDS: Towards an efficient real-time automotive intrusion detection system based on deep evolving stream clustering[J]. Future Generation Computer Systems, 2023, 140: 266-281.
    [8] Wu L, Yuan L, Zhao G, et al. Deep clustering and visualization for end-to-end high-dimensional data analysis[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022, 34(11): 8543-8554.
    [9] Liu T, Luo R, Xu L, et al. Spatial channel attention for deep convolutional neural networks[J]. Mathematics, 2022, 10(10): 1750.
    [10] Lei Y, Yang B, Jiang X, et al. Applications of machine learning to machine fault diagnosis: A review and roadmap[J]. Mechanical Systems and Signal Processing, 2020, 138: 106587.
    [11] 胡斌皓,张建朋,陈鸿昶.基于生成式对抗网络和正类无标签学习的知识图谱补全算法[J].计算机科学,2024,51(01):310-315.Hu B H,Zhang J P ,Chen H X .Knowledge graph completion algorithm based on generative adversarial networks and positive-class semi-supervised learning[J]. Computer Science,2024,51(01):310-315.
    [12] Sharma A, Singh P K, Chandra R. SMOTified-GAN for class imbalanced pattern classification problems[J]. IEEE Access, 2022, 10: 30655-30665.
    [13] 王恒泓,王激尧,徐炜,等.基于领域对抗网络的永磁同步电机初始位置估计[J/OL].电工技术学报,1-14[2024-08- 10].https://doi.org/10.19595/j.cnki.1000-6753.tces.232019.Wang H H,Wang J Y ,Xu W,et al. Initial position estimation of permanent magnet synchronous motors based on domain adversarial networks[J/OL]. Transactions of China Electrotechnical Society, 1-14 [2024-08-10].https://doi.org/10.19595/j.cnki.1000-6753.tces.232019.
    [14] 董绍江,蒋明佑,罗召霞.基于样本扩充与IDANN的刀具状态识别方法[J].重庆大学学报,2023,46(01):16-26.Dong S J, Jiang M Y ,Luo Z X .Tool condition monitoring method based on sample expansion and IDANN.[J].Journal of Chongqing University,2023,46(01):16-26.
    [15] Leevy J L,Khoshgoftaar T M. A survey and analysis of intrusion detection models based on cse-cic-ids2018 big data[J]. Journal of Big Data, 2020, 7: 1-19.
    [16] 姚纤纤,夏鸿斌,刘渊.SDN环境中基于注意力机制的DDoS检测和缓解[J/OL].小型微型计算机系统,1-9[2024-08-09].http://kns.cnki.net/kcms/detail/21.1106.TP.20240530.1039.008.html.Yao Q Q,Xia H B,Liu Y. SDNDDoS detection and mitigation based on attention mechanisms in SDN environments [J/OL].Journal of Chinese Comput Systems, 1-9[2024-08-09].http://kns.cnki.net/kcms/detail/21.1106.TP.20240530.1039.008.html.
    [17] 张子宣, 宗学军, 何戡, 等. 基于CVAE-CatBoost的工业控制网络异常流量检测研究[J].计算机工程,2023,49(05):173-180.Zhang Z X, Zong X J, He K, et al. Research on Industrial Control Network Anomaly Traffic Detection Based on CVAE-CatBoost[J]. Computer Engineering,2023,49(05):173-180.
    [18] Zheng P, Yuan S, Wu X, et al. One-class adversarial nets for fraud detection[C]//2019 Proceedings of the AAAI Conference on Artificial Intelligence.Hawaii, USA:AAAI,2019: 1286-1293.
    [19] 周杭,蒋瑜.基于高对比度子空间的改进孤立森林方法[J].计算机应用研究,2023,40(02):388-393.Zhou H,Jiang Y.Improved isolation forest method based on High-Contrast subspace[J].Application Research of Computers,2023,40(02):388-393.
    [20] 荆志朋, 柴林杰, 胡诗尧.基于改进LSTM-VAE的配电网异常负荷检测方法研究[J/OL].电测与仪表,1-7[2024-08-22].http://kns.cnki.net/kcms/detail/23.1202.th.20240717.1539.002.html. Jing Z P, Chai L J, Hu S Y. The Research on Distribution Network Abnormal Load Detection Method Based on Improved LSTM-VAE[J/OL]. Electrical Measurement Instrumentation,1-7[2024-08-22].http://kns.cnki.net/kcms/detail/23.1202.th.20240717.1539.002.html.
    [21] Leevy J L, Khoshgoftaar T M. A survey and analysis of intrusion detection models based on cse-cic-ids2018 big data[J]. Journal of Big Data, 2020, 7: 1-19.
    [22] Mao W, Wang G, Kou L, et al. Deep domain-adversarial anomaly detection with one-class transfer learning[J]. IEEE/CAA Journal of Automatica Sinica, 2023, 10(2): 524-546.
    [23] Yang Z, Soltani I, Darve E. Anomaly detection with domain adaptation[C]//2023 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, Canada: IEEE,2023: 2958-2967.
    [24] Ruff L, Vandermeulen R, Goernitz N, et al. Deep one-class classification[C]//2018 International conference on machine learning. Stockholm, Sweden: PMLR, 2018: 4393-4402.
    [25] Bergmann P, Fauser M, Sattlegger D, et al. Uninformed students: Student-teacher anomaly detection with discriminative latent embeddings[C]//2020 Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. Seattle, WA, USA: IEEE, 2020: 4183-4192.
    相似文献
    引证文献
    引证文献 [0] 您输入的地址无效!
    没有找到您想要的资源,您输入的路径无效!

    网友评论
    网友评论
    分享到微博
    发 布
引用本文
分享
文章指标
  • 点击次数:135
  • 下载次数: 0
  • HTML阅读次数: 0
  • 引用次数: 0
历史
  • 收稿日期:2024-09-10
  • 最后修改日期:2024-09-10
  • 录用日期:2024-11-11
文章二维码