基于深度时序卷积嵌入聚类的电梯电流数据挖掘

  目前，电梯事故时有发生，而电梯设备的检测往往采用侵入式检测，需停机且难以检测出隐患信息。在先行研究中，在电梯设备出现故障前，其相关电流信号会表现出异常。针对这种情况，本文通过深度聚类的方式，对电流信号进行数据挖掘，以达到提前排查故障隐患的目的。

  传统特征提取和聚类算法往往具有时间复杂度高，鲁棒性低、较依赖于先验知识，以及难以应用到高维数据等缺陷。而对于时序数据聚类，目前也缺乏较为有针对性的特征提取以及聚类算法。本文针对时序数据维度较高，且具有时序之间的因果逻辑关系、数据集往往不平衡的特点，联合时序卷积网络（Temporal Convolutional Network，TCN）和基于局部结构保留的深度嵌入聚类（Improved Deep Embedded Clustering，IDEC）模型，提出一种深度时序卷积嵌入聚类（Deep Temporal Convolutional Embedded Clustering，DTCEC）的方法，挖掘时间序列的隐藏信息，以此来对缺乏标签的数据集进行聚类的目的。首先在UCR数据集上进行实验，通过对比选择最佳参数，并且使用传统的聚类方法进行实验。结果显示，在多个数据集上，DTCEC在聚类精确度、归一化互信息以及F1-score都有着良好的表现，相比于传统的聚类模型，取得了更好的效果。

  本文将DTCEC应用于电梯电流数据，将电流信号进行预处理后，使用DTCEC对电梯的多种电流信号进行聚类，以此来对各种信号的不同类别进行分析。由于标签缺乏，本文使用聚类内部评价指标对聚类结果进行评判，并且使用IDEC以及时序卷积自编码器（Encoder-Decoder Temporal Convolutional Network，ED-TCN）算法进行对比，证明了时序卷积以及IDEC模块在聚类方面的优势。另外，本文针对电梯电流数据集数据海量且极不平衡的特点，对DTCEC模型进行改进，建立了DTCEC联合自编码器-K均值聚类（AE_K-means）异常检测算法，对极少比例的异常信号进行挖掘，从而获取异常样本，为后续研究提供充分的数据标签以及依据。

[1] 市场监督管理总局关于 2019 年全国特种设备安全状况的通告[J]. 中国特种设备安全, 2020, 36(04): 1-4.
[2] 机电工程 署 ： 已 报 告 的 升 降 机 意 外 记 录 (799)[EB/OL]. https://www.emsd.gov.hk/tc/lifts_and_escalators_safety/publications/contractors_performance_rating/reported_lift_incident_records/index.html.
[3] Wang C, Zhang R, Zhang Q. Analysis of transverse vibration acceleration for a high -speed elevator with random parameter based on perturbation theory[J]. Int. J. Acoustics Vibration, 2017, 22: 218-223.
[4] Chai S, Li X I, Jia Y, et al. A Non-Intrusive Deep Learning Based Diagnosis System for Elevators[J]. IEEE Access, 2021, 9: 20993-21003.
[5] 李明亮. 时间序列相似性聚类算法研究[D]. 湖南大学, 2010.
[6] Min E, Guo X, Liu Q, et al. A survey of clustering with deep learning: From the perspective of network architecture[J]. IEEE Access, 2018, 6: 39501-39514.
[7] Soleimany G, Abessi M. A New Similarity Measure for Time Series Data Mining Based on Longest Common Subsequence[J]. American Journal of Data Mining and Knowledge Discovery, 2019, 4(1): 32.
[8] Agarwal A, Maheswaran R, Marwan N, et al. Wavelet-based multiscale similarity measure for complex networks[J]. The European Physical Journal B, 2018, 91(11): 1 -12.
[9] Radhakrishna V, Aljawarneh S A, Kumar P V, et al. A novel fuzzy similarity measure and prevalence estimation approach for similarity profiled temporal association pattern mining[J]. Future Generation Computer Systems, 2018, 83: 582-595.
[10] Zhang M, Pi D. A new time series representation model and corresponding similarity measure for fast and accurate similarity detection[J]. IEEE Access, 2017, 5: 24503-24519.
[11] 甄远婷, 冶继民, 李国荣. 基于中心 Copula 函数相似性度量的时间序列聚类方法[J]. 陕西师范大学学报(自然科学版), 2021, 49(01): 29-36.
[12] 夏寒松, 张力生, 桑春艳. 基于 LDTW 的动态时间规整改进算法[J]. 计算机工程: 1-14.
[13] 李国荣, 冶继民, 甄远婷. 基于新的鲁棒相似性度量的时间序列聚类[J]. 计算机应用: 1-7.
[14] Soni A N. Feature Extraction Methods for Time Series Functions using Machine Learning[J]. International Journal of Innovative Research in Science, Engineering and Technology, 2018, 7(8): 8661-8665.
[15] Bauer A, Züfle M, Herbst N, et al. Telescope: An automatic feature extraction and transformation approach for time series forecasting on a level-playing field[C]. 2020 IEEE 36th International Conference on Data Engineering (ICDE), 202 0: 1902-1905.
[16] 陈海兰. 面向聚类及预测的时间序列信息粒化方法研究[D]. 北京:北京科技大学, 2021.
[17] 刘学. 基于流形学习的时间序列聚类研究[D]. 石家庄:河北经贸大学, 2016.
[18] Han M, Feng S, Chen C P, et al. Structured manifold broad learning system: A manifold perspective for large-scale chaotic time series analysis and prediction[J]. IEEE Transactions on Knowledge and Data Engineering, 2018, 31(9): 1809 -1821.
[19] Li Y, Li Y, Chen X, et al. Denoising and feature extraction algorithms using NPE combined with VMD and their applications in ship-radiated noise[J]. Symmetry, 2017, 9(11): 256.
[20] Lapidot I, Bonastre J-F. Generalized viterbi-based models for time-series segmentation applied to speaker diarization[C]. ODYSSEY 2012-The Speaker and Language Recognition Workshop, 2012.
[21] Zang D, Liu J, Wang H. Markov chain-based feature extraction for anomaly detection in time series and its industrial application[C]. 2018 Chinese Control And Decision Conference (CCDC), 2018: 1059-1063.
[22] Yang Y, Jiang J. Adaptive bi-weighting toward automatic initialization and model selection for HMM-based hybrid meta-clustering ensembles[J]. IEEE Transactions on Cybernetics, 2018, 49(5): 1657-1668.
[23] Jung Y, Park J. Scalable Hybrid Hidden Markov Model with Gaussian Process Emission for Sequential Time-series Observations[C]. Third Symposium on Advances in Approximate Bayesian Inference, 2020.
[24] Kanaan M, Benabdeslem K, Kheddouci H. A Generative Time Series Clustering Framework Based on an Ensemble Mixture of HMMs[C]. 2020 IEEE 32nd International Conference on Tools with Artificial Intelligence (ICTAI), 2020: 793-798.
[25] Schmidhuber J. Deep learning in neural networks: An overview[J]. Neural Networks, 2015, 61: 85-117.
[26] Yang B, Fu X, Sidiropoulos N D, et al. Towards k-means-friendly spaces: Simultaneous deep learning and clustering[C]. International Conference on Machine Learning, 2017: 3861-3870.
[27] Huang P, Huang Y, Wang W, et al. Deep embedding network for clustering[C]. 2014 22nd International conference on pattern recognition, 2014: 1532-1537.
[28] Chen D, Lv J, Zhang Y. Unsupervised multi-manifold clustering by learning deep representation[C]. Workshops at the thirty-first AAAI conference on artificial intelligence, 2017.
[29] Shah S A, Koltun V. Deep continuous clustering[J]. arXiv preprint arXiv:1803.01449, 2018.
[30] Arif M, Wang G, Chen S. Deep learning with non-parametric regression model for traffic flow prediction[C]. 2018 IEEE 16th Intl Conf on Dependable, Autonomic and Secure Computing, 16th Intl Conf on Pervasive Intelligence and Computing, 4th Intl Conf on Big Data Intelligence and Computing and Cyber Science and Technology Congress (DASC/PiCom/DataCom/CyberSciTech), 2018: 681-688.
[31] Xie J, Girshick R, Farhadi A. Unsupervised deep embedding for clustering analysis[C]. International conference on machine learning, 2016: 478-487.
[32] Li F, Qiao H, Zhang B. Discriminatively Boosted Image Clustering with Fully Convolutional Auto-Encoders[J]. Pattern Recognition, 2017, 83: 161-173.
[33] Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[J]. Advances in neural information processing systems, 2014, 27.
[34] Harchaoui W, Mattei P-A, Bouveyron C. Deep adversarial Gaussian mixture auto encoder for clustering[J], 2017.
[35] Dosovitskiy A, Springenberg J T, Riedmiller M, et al. Discriminative unsupervised feature learning with convolutional neural networks[J]. Advances in neural information processing systems, 2014, 27.
[36] Chen X, Duan Y, Houthooft R, et al. Infogan: Interpretable representation learning by information maximizing generative adversarial nets[C]. Proceedings of the 30th International Conference on Neural Information Processing Systems, 2016: 2180 -2188.
[37] Jiang Z, Zheng Y, Tan H, et al.: Variational deep embedding: A generative approach to clustering, CoRR, 2016.
[38] Liao W, Guo Y, Chen X, et al. A unified unsupervised gaussian mixture variational autoencoder for high dimensional outlier detection[C]. 2018 IEEE International Conference on Big Data (Big Data), 2018: 1208-1217.
[39] Bai S, Kolter J Z, Koltun V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[J]. arXiv preprint arXiv:1803.01271, 2018.
[40] Zeiler M D, Taylor G W, Fergus R. Adaptive deconvolutional networks for mid and high level feature learning[C]. 2011 international conference on computer vision, 2011: 2018-2025.
[41] Zeiler M D, Fergus R. Visualizing and understanding convolutional networks[C]. European conference on computer vision, 2014: 818-833.
[42] Lea C, Flynn M D, Vidal R, et al. Temporal convolutional networks for action segmentation and detection[C]. proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 156-165.
[43] Van Den Oord A, Dieleman S, Zen H, et al. WaveNet: A generative model for raw audio[J]. SSW, 2016, 125: 2.
[44] Bengio, Yoshua, Courville, et al. Representation Learning: A Review and New Perspectives[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2013, 35(8): 1798-1828.
[45] Guo X, Gao L, Liu X, et al. Improved deep embedded clustering with local structure preservation[C]. Ijcai, 2017: 1753-1759.
[46] Amarbayasgalan T, Jargalsaikhan B, Ryu K H. Unsupervised novelty detection using deep autoencoders with density based clustering[J]. Applied Sciences, 2018, 8(9): 1468.
[47] Zhang C, Song D, Chen Y, et al. A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data[J], 2018.