Unsupervised Time Series Segmentation: A Survey on Recent Advances
Source abstract
Abstract
Time series segmentation has attracted more interests in recent years, which aims to segment time series into different segments, each reflects a state of the monitored objects. Although there have been many surveys on time series segmentation, most of them focus more on change point detection (CPD) methods and overlook the advances in boundary detection (BD) and state detection (SD) methods. In this paper, we categorize time series segmentation methods into CPD, BD, and SD methods, with a specific focus on recent advances in BD and SD methods. Within the scope of BD and SD, we subdivide the methods based on their underlying models/techniques and focus on the milestones that have shaped the development trajectory of each category. As a conclusion, we found that: (1) Existing methods failed to provide sufficient support for online working, with only a few methods supporting online deployment; (2) Most existing methods require the specification of parameters, which hinders their ability to work adaptively; (3) Existing SD methods do not attach importance to accurate detection of boundary points in evaluation, which may lead to limitations in boundary point detection. We highlight the ability to working online and adaptively as important attributes of segmentation methods, the boundary detection accuracy as a neglected metrics for SD methods.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Time Series Analysis and Forecasting · Anomaly Detection Techniques and Applications · Complex Systems and Time Series Analysis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (68)
Show source-provided references
1. Y. Matsubara, Y. Sakurai, and C. Faloutsos, “AutoPlait: Automatic mining of co-evolving time sequences,” presented at the ACM SIGMOD Int. Conf. Manag. Data, New York, NY, USA, 2014. doi: 10.1145/2588555.2588556.10.1145/2588555.2588556 ↗
2. Time2State: An unsupervised framework for inferring the latent states in time series data
3. D. Hallac, S. Vare, S. Boyd, and J. Leskovec, “Toeplitz inverse covariance-based clustering of multivariate time series data,” presented at the 23rd ACM SIGKDD Int. Conf. Knowl. Discov. Data Mining, New York, NY, USA, 2017. doi: 10.1145/3097983.3098060.10.1145/3097983.3098060 ↗
4. ESPRESSO: Entropy and shape aware time-series segmentation for processing heterogeneous sensor data10.1145/3411832 ↗
5. Domain agnostic online semantic segmentation for multi-dimensional time series10.1007/s10618-018-0589-3 ↗
6. A survey of methods for time series change point detection10.1007/s10115-016-0987-z ↗
7. A survey on change detection and time series analysis with applications10.3390/app11136141 ↗
8. Selective review of offline change point detection methods10.1016/j.sigpro.2019.107299 ↗
9. P. Schäfer, A. Ermshaus, and U. Leser, “ClaSP—time series segmentation,” presented at the 30th ACM Int. Conf. Inf. Knowl. Manag., New York, NY, USA, 2021, pp. 1578–1587. doi: 10.1145/3459637.3482240.10.1145/3459637.3482240 ↗
10. J. Harguess and J. Aggarwal, “Semantic labeling of track events using time series segmentation and shape analysis,” presented at the 16th IEEE Int. Conf. Image Process. (ICIP), Cairo, Egypt, 2009, pp. 4317–4320. doi: 10.1109/ICIP.2009.5413671.10.1109/ICIP.2009.5413671 ↗
11. P. Wang, H. Wang, and W. Wang, “Finding semantics in time series,” presented at the ACM SIGMOD Int. Conf. Manage. Data, Athens, Greece, 2011, pp. 385–396. doi: 10.1145/1989323.1989364.10.1145/1989323.1989364 ↗
12. A fast and robust method for detecting trend turning points in InSAR displacement time series10.1016/j.cageo.2024.105546 ↗
13. BFAST lite: A lightweight break detection method for time series analysis10.3390/rs13163308 ↗
14. E. Ghaderpour et al., “Least-squares wavelet analysis of rainfalls and landslide displacement time series derived by PS-InSAR,” presented at the Int. Conf. Time Ser. Forecast., Gran, Canaria, Spain, 2022, 117–132. doi: 10.1007/978-3-031-40209-8_9.10.1007/978-3-031-40209-8_9 ↗
15. Least-squares wavelet analysis of unequally spaced and non-stationary time series and its applications10.1007/s11004-017-9691-0 ↗
16. Continuous and discrete wavelet transforms10.1137/1031129 ↗
17. Fast algorithms for discrete and continuous wavelet transforms10.1109/18.119724 ↗
18. On empirical mode decomposition and its algorithms
19. A. Zeiler, R. Faltermeier, I. Keck, A. Tomé, C. Puntonet and E. Lang, “Empirical mode decomposition-an introduction,” presented at the Int. Joint Conf. Neural Netw. (IJCNN), Barcelona, Spain, 2010, pp. 1–8.10.1109/IJCNN.2010.5596829 ↗
20. Movement primitive segmentation for human motion modeling: A framework for analysis10.1109/THMS.2015.2493536 ↗
21. S. Gharghabi, Y. Ding, C. Yeh, K. Kamgar, L. Ulanova and E. Keogh, “Matrix profile VIII: Domain agnostic online semantic segmentation at superhuman performance levels,” presented at the 2017 IEEE Int. Conf. Data Mining (ICDM), New Orleans, LA, USA, 2017, pp. 117–126. doi: 10.1109/ICDM.2017.21.10.1109/ICDM.2017.21 ↗
22. J. Heinisch, C. Anderson, and K. David, “Angry or climbing stairs? Towards physiological emotion recognition in the wild,” presented at the 2019 IEEE Int. Conf. Pervasive Comput. Commun. Workshops (PerCom Workshops), Kyoto, Japan, 2019, pp. 486–491.10.1109/PERCOMW.2019.8730725 ↗
23. A first step towards eye state prediction using eeg
24. MoCap. 2024. Accessed: Jul. 3, 2024. [Online]. Available: http://mocap.cs.cmu.edu
25. A tutorial on spectral clustering10.1007/s11222-007-9033-z ↗
26. A. Reiss and D. Stricker, “Introducing a new benchmarked dataset for activity monitoring,” presented at the 16th Int. Symp. Wearable Comput., Newcastle, UK, 2012, pp. 108–109. doi: 10.1109/ISWC.2012.13.10.1109/ISWC.2012.13 ↗
27. M. Zhang and A. Sawchuk, “USC-HAD: A daily activity dataset for ubiquitous activity recognition using wearable sensors,” presented at the ACM Conf. Ubiquitous Comput., New York, NY, USA, 2012, pp. 1036–1043. doi: 10.1145/2370216.2370438.10.1145/2370216.2370438 ↗
28. D. Roggen et al., “Collecting complex activity datasets in highly rich networked sensor environments,” Presented at the 2010 Seventh Int. Conf. Netw. Sens. Syst. (INSS), Kassel, Germany, 2010, pp. 233–240.10.1109/INSS.2010.5573462 ↗
29. Distributed recognition of human actions using wearable motion sensor networks10.3233/AIS-2009-0016 ↗
30. E. Tapia, S. Intille, L. Lopez, and K. Larson, “The design of a portable kit of wireless sensors for naturalistic data collection,” presented at the Pervasive Comput. 4th Int. Conf., Dublin, Ireland, 2006, pp. 117–134.10.1007/11748625_8 ↗
31. F. De la Torre et al., “Guide to the carnegie mellon university multimodal activity CMU-MMAC database,” 2024. Accessed: Jul. 3, 2024. [Online]. Available: https://www.ri.cmu.edu/publications/guide-to-the-carnegie-mellon-university-multimodal-activity-cmu-mmac-database/
32. TSAGen: Synthetic time series generation for KPI anomaly detection10.1109/TNSM.2021.3098784 ↗
33. Information gain-based metric for recognizing transitions in human activities10.1016/j.pmcj.2017.01.003 ↗
34. Hidden markov models10.1016/S0959-440X(96)80056-X ↗
35. K. Murphy, “Hidden semi-markov models (HSMMs),” 2002. Accessed: Jul. 3, 2024. [Online]. Available: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=a2125e4fba6c69ff7a6d247bcf04312bb2b79606
36. Bayesian nonparametric hidden semi-markov models
37. Z. Lai, H. Li, D. Zhang, Y. Zhao, W. Qian and C. Jensen, “E2Usd: Efficient-yet-effective unsupervised state detection for multivariate time series,” presented at the ACM on Web Conf. 2024, New York, NY, USA, 2024, pp. 3010–3021. doi: 10.1145/3589334.3645593.10.1145/3589334.3645593 ↗
38. StreamScope: Automatic pattern discovery over data streams
39. Clifford · Disorder in Physical Systems: A Volume in Honour of John M. Hammersley · 1990 · 19
40. What is a hidden Markov model?10.1038/nbt1004-1315 ↗
41. Benedetto · Principles of Digital Transmission: With Wireless Applications · 2002 · 807
42. What is the expectation maximization algorithm?10.1038/nbt1406 ↗
43. Hidden semi-Markov models10.1016/j.artint.2009.11.011 ↗
44. The infinite hidden Markov model
45. K. Murphy, “Dynamic bayesian networks,” 2002. Accessed: Jul. 3, 2024. [Online]. Available: https://webdocs.cs.ualberta.ca/~rgreiner/C-366/RG-2002-SLIDES/dbn-murphy.pdf
46. E. Fox, E. Sudderth, M. Jordan, and A. Willsky, “An HDP-HMM for systems with state persistence,” presented at the 25th Int. Conf. Mach. Learn., Helsinki, Finland, 2008, pp. 312–319.10.1145/1390156.1390196 ↗
47. Sharing clusters among related groups: Hierarchical dirichlet processes
48. Hierarchical dirichlet processes10.1198/016214506000000302 ↗
49. M. Ankerst, M. Breunig, H. Kriegel, and J. Sander, “OPTICS: Ordering points to identify the clustering structure,” presented at the ACM SIGMOD Int. Conf. Manage. Data, New York, NY, USA, 1999, pp. 49–60. doi: 10.1145/304182.304187.10.1145/304182.304187 ↗
50. M. Ester, H. Kriegel, J. Sander, and X. Xu, “A density-based algorithm for discovering clusters in large spatial databases with noise,” presented at the ACM SIGKDD Conf. Knowl. Discov. Data Mining, Portland, Oregon, 1996, pp. 226–231.
51. K-means clustering: A half-century synthesis10.1348/000711005X48266 ↗
52. Optimal neighborhood multiple kernel clustering with adaptive local kernels
53. Fast parameter-free multi-view subspace clustering with consensus anchor guidance10.1109/TIP.2021.3131941 ↗
54. HVGH: Unsupervised segmentation for high-dimensional time series using deep neural compression and statistical generative model
55. E. Eldele et al., “Time-series representation learning via temporal and contextual contrasting,” presented at the Thirtieth Int. Joint Conf. Artif. Intell. (IJCAI), 2021, pp. 2352–2359.10.24963/ijcai.2021/324 ↗
56. Unsupervised scalable representation learning for multivariate time series
57. A. Van den Oord, Y. Li, and O. Vinyals, “Representation learning with contrastive predictive coding,” arXiv preprint arXiv:1807.03748, 2018.
58. S. Tonekaboni, D. Eytan, and A. Goldenberg, “Unsupervised representation learning for time series with temporal neighborhood coding,” arXiv preprint arXiv:2106.00750, 2021.
59. Z. Yue et al., “Ts2Vec: Towards universal representation of time series,” presented at the AAAI Conf. Artif. Intell., 2022, pp. 8980–8987.10.1609/aaai.v36i8.20881 ↗
60. In pursuit of beauty: Aesthetic-aware and context-adaptive photo selection in crowdsensing10.1109/TKDE.2023.3237969 ↗
61. RayMVSNet++: Learning ray-based 1D implicit fields for accurate multi-view stereo
62. The minimum description length principle in coding and modeling10.1109/18.720554 ↗
63. N. Tatti and J. Vreeken, “The long and the short of it: Summarising event sequences with serial episodes,” presented at the 18th ACM SIGKDD Int. Conf. Knowl. Discov. Data Mining, Beijing, China, 2012, pp. 462–470. doi: 10.1145/2339530.2339606.10.1145/2339530.2339606 ↗
64. K. Kawabata, Y. Matsubara, and Y. Sakurai, “Automatic sequential pattern mining in data streams,” presented at the 28th ACM Int. Conf. Inf. Knowl. Manage., Beijing, China, 2019, pp. 1733–1742. doi: 10.1145/3357384.3358002.10.1145/3357384.3358002 ↗
65. Y. He, X. Chu, and Y. Wang, “Neighbor profile: Bagging nearest neighbors for unsupervised time series mining,” presented at the IEEE 36th Int. Conf. Data Eng. (ICDE), Dallas, TX, USA, 2020, pp. 373–384. doi: 10.1109/ICDE48307.2020.00039.10.1109/ICDE48307.2020.00039 ↗
66. F. Madrid, S. Imani, R. Mercer, Z. Zimmerman, N. Shakibay and E. Keogh, “Matrix profile XX: Finding and visualizing time series motifs of all lengths using the matrix profile,” presented at the IEEE Int. Conf. Big Knowl. (ICBK), Beijing, China, 2019, pp. 175–182. doi: 10.1109/ICBK.2019.00031.10.1109/ICBK.2019.00031 ↗
67. C. Yeh, N. Kavantzas, and E. Keogh, “Matrix profile VI: Meaningful multidimensional motif discovery,” presented at the IEEE Int. Conf. Data Mining (ICDM), Beijing, China, 2017, pp. 565–574. doi: 10.1109/ICDM.2017.66.10.1109/ICDM.2017.66 ↗
68. C. Yeh et al., “Matrix profile I: All pairs similarity joins for time series: a unifying view that includes motifs, discords and shapelets,” presented at the IEEE 16th Int. Conf. Data Mining (ICDM), Barcelona, Spain, 2016, pp. 1317–1322. doi: 10.1109/ICDM.2016.0179.10.1109/ICDM.2016.0179 ↗