计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (6A): 220700094-7.doi: 10.11896/jsjkx.220700094

• 大数据&数据科学 • 上一篇    下一篇

基于多模态特征融合的时间序列异常检测

张国华, 燕雪峰, 关东海   

  1. 南京航空航天大学计算机科学与技术学院软件新技术与产业化协同创新中心 南京 211106
  • 出版日期:2023-06-10 发布日期:2023-06-12
  • 通讯作者: 燕雪峰(yxf@nuaa.edu.cn)
  • 作者简介:(zgh2020@nuaa.edu.cn)

Anomaly Detection of Time-series Based on Multi-modal Feature Fusion

ZHANG Guohua, YAN Xuefeng, GUAN Donghai   

  1. College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Software New Technology and Industrialization Collaborative Innovation Center,Nanjing 211106,China
  • Online:2023-06-10 Published:2023-06-12
  • About author:ZHANG Guohua,born in 1995,postgraduate.His main research interests include machine learning and anomaly detection. YAN Xuefeng,born in 1975,Ph.D,professor,is a member of China Computer Federation.His main research interests include intelligent computing,model-based systems engineering,simulation,and evaluation.

PDF (PC)

361

摘要： 多元时间序列的有效异常检测对于数据的分析挖掘具有重要意义。然而,已有的检测方法大多基于单模态,不能有效利用时间序列在多模态空间中的分布信息,对于多模态特征缺乏自适应融合方式且难以提取其时空依赖关系。为此,提出了一种多模态特征融合的时间序列异常检测方法,建立了一个多模态特征自适应融合模块,通过一维卷积网络和软选择方式对多元时间序列的多模态特征进行自适应融合。对于融合后的多模态特征,构建由时间注意力和空间注意力组成的时空注意力模块,同时提取其时间和空间依赖关系得到时空注意力向量,由时空注意力向量得到模型预测结果。通过学习正常样本分布,根据预测值与真实值的误差度量实现异常检测。在4个公开数据集上进行测试,结果表明,所提方法优于其他模型,证明了所提方法的有效性。

关键词: 时间序列, 异常检测, 长短期记忆网络, 一维卷积神经网络, 注意力机制

Abstract: Effective anomaly detection of multivariate time series is important for data mining analysis.However,most of the exi-sting detection methods are based on single modality,they cannot effectively utilize the distribution information of time series in multi-modal space.For multi-modal features,there is no effective adaptive fusion method and extraction method of spatial-temporal dependence.In this paper,a time series anomaly detection method based on multi-modal feature fusion is proposed.The multi-modal feature adaptive fusion module is established,it can adaptively fuse the multi-modal features through convolution network and soft selection mode.The spatial-temporal attention module is proposed,it is composed of temporal attention and spatial attention.It extracts spatial-temporal dependence of the multi-modal features and outputs the spatial-temporal attention vector.Then the model prediction results are obtained based on the spatial-temporal attention vector.By learning the distribution of normal samples,anomaly detection result is obtained according to the error measure between the predicted values and the real values.The proposed method is compared with other state-of-the-art models on four public datasets,and results demonstrate its effectiveness.

Key words: Time series, Anomaly detection, LSTM, 1D-CNN, Attention mechanism

中图分类号: 

  • TP391
[1]GOTHWAL H,KEDAWAT S,KUMAR R.Cardiac arrhythmias detection in an ECG beat signal using fast fourier transform and artificial neural network[J].Journal of Biomedical Science and Engineering,2011,4(4):289-296.
[2]KIRANYAZ S,INCE T,GABBOUJ M.Real-time patient-spe-cific ECG classification by 1-D convolutional neural networks[J].IEEE Transactions on Biomedical Engineering,2015,63(3):664-675.
[3]SU Y,ZHAO Y,NIU C,et al.Robust anomaly detection for multivariate time series through stochastic recurrent neural network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2019:2828-2837.
[4]HUNDMAN K,CONSTANTINOU V,LAPORTE C,et al.Detecting spacecraft anomalies using lstms and nonparametric dynamic thresholding[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2018:387-395.
[5]ZHANG Y,CHEN Y,WANG J,et al.Unsupervised deep ano-maly detection for multi-sensor time-series signals[J].IEEE Transactions on Knowledge and Data Engineering,2023,35(2):2118-2132.
[6]MALHOTRA P,RAMAKRISHNAN A,ANAND G,et al.LSTM-based encoder-decoder for multi-sensor anomaly detection[J].arXiv:1607.00148,2016.
[7]AUDIBERT J,MICHIARDI P,GUYARD F,et al.USAD:unsupervised anomaly detection on multivariate time series[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2020:3395-3404.
[8]DU B,SUN X,YE J,et al.GAN-Based Anomaly Detection for Multivariate Time Series Using Polluted Training Set[J].IEEE Transactions on Knowledge and Data Engineering,2021,30(3):1028-1037.
[9]LI S,WEN J.A model-based fault detection and diagnostic methodology based on PCA method and wavelet transform[J].Energy and Buildings,2014,68:63-71.
[10]LIU S,ZHOU B,DING Q,et al.Time Series Anomaly Detection with Adversarial Reconstruction Networks[J].IEEE Transactions on Knowledge and Data Engineering,2023,35(4):4293-4306.
[11]GARG A,ZHANG W,SAMARAN J,et al.An Evaluation ofAnomaly Detection and Diagnosis in Multivariate Time Series[J].IEEE Transactions on Neural Networks and Learning Systems,2022,33(6):2508-2517.
[12]LIN S,CLARK R,BIRKE R,et al.Anomaly detection for time series using vae-lstm hybrid model[C]//2020 IEEE International Conference on Acoustics,Speech and Signal Processing(ICASSP 2020).IEEE,2020:4322-4326.
[13]DING N,GAO H,BU H,et al.Multivariate-time-series-driven real-time anomaly detection based on bayesian network[J].Sensors,2018,18(10):3367.
[14]GUGULOTHU N,MALHOTRA P,VIG L,et al.Sparse neural networks for anomaly detection in high-dimensional time series[C]//AI4IOT workshop in Conjunction with ICML,IJCAI and ECAI.2018:1551-3203.
[15]LAI G,CHANG W C,YANG Y,et al.Modeling long-and short-term temporal patterns with deep neural networks[C]//The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval.2018:95-104.
[16]SHIH S Y,SUN F K,LEE H.Temporal pattern attention for multivariate time series forecasting[J].Machine Learning,2019,108(8):1421-1441.
[17]DAI Y,GIESEKE F,OEHMCKE S,et al.Attentional feature fusion[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision.2021:3560-3569.
[18]LI X,WANG W,HU X,et al.Selective kernel networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2019:510-519.
[19]KONG Z Q,DENG L,TANG B P,et al.Fault diagnosis of planetary gearbox based on deep learning with time-frequency fusion and attention mechanism[J].Chinese Journal of Scientific Instrument,2019,40(6):221-227.
[20]HUANG X,ZHANG F,FAN H,et al.Multimodal Adversarial Learning Based Unsupervised Time Series Anomaly Detection[J].Journal of Computer Research and Development,2021,58(8):1655.
[21]COCCONCELLI M,ZIMROZ R,RUBINI R,et al.STFT based approach for ball bearing fault detection in a varying speed motor[M]//Condition Monitoring of Machinery in Non-stationary Operations.Berlin:Springer,2012:41-50.
[22]WU R,KEOGH E.Current time series anomaly detectionbenchmarks are flawed and are creating the illusion of progress[J].IEEE Transactions on Knowledge and Data Engineering,2022:1479-1480.
[23]KEOGH E,LIN J,FU A.Hot sax:Efficiently finding the most unusual time series subsequence[C]//Fifth IEEE International Conference on Data Mining,2005:226-233.
[24]RAJPURKAR P,HANNUN A Y,HAGHPANAHI M,et al.Cardiologist-level arrhythmia detection with convolutional neural networks[J].arXiv:1707.01836,2017.
[25]SAEEDI EMADI H,MAZINANI S M.A novel anomaly detection algorithm using DBSCAN and SVM in wireless sensor networks[J].Wireless Personal Communications,2018,98(2):2025-2035.
[26]HARISH B S,KUMAR S V A.Anomaly based intrusion detection using modified fuzzy clustering[J].Int. J. Interact. Multim. Artif. Intell.,2017,4(6):54-59.
[27]SIFFER A,FOUQUE P A,TERMIER A,et al.Anomaly detection in streams with extreme value theory[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2017:1067-1075.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发