(School of Electrical Engineering and Computer Science, Vietnamese–German University, Viet Nam)(2) Khoa Nguyen
(Information and Communication Technology, Chungbuk National University, Korea, Republic of)(3) Taehong Kim
(Information and Communication Technology, Chungbuk National University, Korea, Republic of)(4) Ngoc Hong Tran
(School of Electrical Engineering and Computer Science, Vietnamese–German University, Viet Nam)(5) * Vinh Dinh
(Vietnamese German University,
AI VIETNAM, Viet Nam)*corresponding author
AbstractTime series data analysis is crucial for real-world applications. While deep learning has advanced in this field, it still faces challenges, such as limited or poor-quality data. In areas like computer vision, data augmentation has been widely used and highly effective in addressing similar issues. However, these techniques are not as commonly explored or applied in the time series domain. This paper addresses the gap by evaluating basic data augmentation techniques using MLP, CNN, and Transformer architectures, prioritized for their alignment with state-of-the-art trends in time series analysis rather than traditional RNN-based methods. The goal is to expand the use of data augmentation in time series analysis. The paper proposed EMixup, which adapts the Mixup method from image processing to time series data. This adaptation involves mixing samples while aiming to maintain the data's temporal structure and integrating target contributions into the loss function. Empirical studies show that EMixup improves the performance of time series models across various architectures (improving 23/24 forecasting cases and 12/24 classification cases). It demonstrates broad applicability and strong results in tasks like forecasting and classification, highlighting its potential utility across diverse time series applications.
KeywordsTime series augmentation; Mixed-sample augmentation; Times series forecasting
|
DOIhttps://doi.org/10.26555/ijain.v11i2.1592 |
Article metricsAbstract views : 650 | PDF views : 68 |
Cite |
Full Text Download
|
References
[1] A. Zeng, M. Chen, L. Zhang, and Q. Xu, “Are Transformers Effective for Time Series Forecasting?,” Proc. AAAI Conf. Artif. Intell., vol. 37, no. 9, pp. 11121–11128, Jun. 2023, doi: 10.1609/aaai.v37i9.26317.
[2] H. Wu, T. Hu, Y. Liu, H. Zhou, J. Wang, and M. Long, “TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis,” 11th Int. Conf. Learn. Represent. ICLR 2023, pp. 1–23, Oct. 2022. [Online]. Available at: https://arxiv.org/pdf/2210.02186.
[3] D. Dera, S. Ahmed, N. C. Bouaynaya, and G. Rasool, “TRustworthy Uncertainty Propagation for Sequential Time-Series Analysis in RNNs,” IEEE Trans. Knowl. Data Eng., vol. 36, no. 2, pp. 1–13, Feb. 2023, doi: 10.1109/TKDE.2023.3288628.
[4] Y. Liu et al., “iTransformer: Inverted Transformers Are Effective for Time Series Forecasting,” 12th Int. Conf. Learn. Represent. ICLR 2024, pp. 1–25, Oct. 2023. [Online]. Available at: https://arxiv.org/pdf/2310.06625.
[5] Z. Wang et al., “A Comprehensive Survey on Data Augmentation,” arXiv, pp. 1–20, May 2024. [Online]. Available at: https://arxiv.org/pdf/2405.09591.
[6] A. Lazcano, M. A. Jaramillo-Morán, and J. E. Sandubete, “Back to Basics: The Power of the Multilayer Perceptron in Financial Time Series Forecasting,” Mathematics, vol. 12, no. 12, p. 1920, Jun. 2024, doi: 10.3390/math12121920.
[7] R. Espinosa, F. Jiménez, and J. Palma, “Embedded feature selection in LSTM networks with multi-objective evolutionary ensemble learning for time series forecasting,” arXiv, pp. 1–25, Dec. 2023. [Online]. Available at: https://arxiv.org/pdf/2312.17517.
[8] W. Chen and J. Li, “SIGAN-CNN: Convolutional Neural Network Based Stepwise Improving Generative Adversarial Network for Time Series Classification of Small Sample Size,” IEEE Access, vol. 12, pp. 85499–85510, 2024, doi: 10.1109/ACCESS.2024.3413948.
[9] A. Das, W. Kong, R. Sen, and Y. Zhou, “A decoder-only foundation model for time-series forecasting,” Proc. Mach. Learn. Res., vol. 235, pp. 10148–10167, Oct. 2023. [Online]. Available at: https://arxiv.org/pdf/2310.10688.
[10] V. Naghashi, M. Boukadoum, and A. B. Diallo, “A multiscale model for multivariate time series forecasting,” Sci. Rep., vol. 15, no. 1, p. 1565, Dec. 2025, doi: 10.1038/S41598-024-82417-4.
[11] B. Liu, Z. Zhang, and R. Cui, “Efficient Time Series Augmentation Methods,” in 2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), Oct. 2020, pp. 1004–1009, doi: 10.1109/CISP-BMEI51763.2020.9263602.
[12] W. Yang, J. Yuan, and X. Wang, “SFCC: Data Augmentation with Stratified Fourier Coefficients Combination for Time Series Classification,” Neural Process. Lett., vol. 55, no. 2, pp. 1833–1846, Apr. 2023, doi: 10.1007/s11063-022-10965-9.
[13] M. Goubeaud, N. Gmyrek, F. Ghorban, L. Schelkes, and A. Kummert, “Random Noise Boxes: Data Augmentation for Spectrograms,” in 2021 IEEE International Conference on Progress in Informatics and Computing (PIC), Dec. 2021, pp. 24–28, doi: 10.1109/PIC53636.2021.9687058.
[14] W. Zhao and L. Fan, “Time-series representation learning via Time-Frequency Fusion Contrasting,” Front. Artif. Intell., vol. 7, p. 1414352, Jun. 2024, doi: 10.3389/frai.2024.1414352.
[15] Y. Deng, R. Liang, D. Wang, A. Li, and F. Xiao, “Decomposition-based Data Augmentation for Time-series Building Load Data,” in Proceedings of the 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, Nov. 2023, pp. 51–60, doi: 10.1145/3600100.3623727.
[16] B. Kalina, J.-H. Lee, and K.-T. Na, “Enhancing Portfolio Performance through Financial Time-Series Decomposition-Based Variational Encoder-Decoder Data Augmentation,” Symmetry (Basel)., vol. 16, no. 3, p. 283, Feb. 2024, doi: 10.3390/sym16030283.
[17] X. Wu, D. Zhang, G. Li, X. Gao, B. Metcalfe, and L. Chen, “Data augmentation for invasive brain–computer interfaces based on stereo-electroencephalography (SEEG),” J. Neural Eng., vol. 21, no. 1, p. 016026, Feb. 2024, doi: 10.1088/1741-2552/ad200e.
[18] L. Wang, L. Zeng, and J. Li, “AEC-GAN: Adversarial Error Correction GANs for Auto-Regressive Long Time-Series Generation,” Proc. AAAI Conf. Artif. Intell., vol. 37, no. 8, pp. 10140–10148, Jun. 2023, doi: 10.1609/aaai.v37i8.26208.
[19] H. Zhang, M. Cisse, Y. N. Dauphin, and D. Lopez-Paz, “Mixup: Beyond Empirical Risk Minimization,” 6th Int. Conf. Learn. Represent. ICLR 2018 - Conf. Track Proc., pp. 1–13, Oct. 2017, 2025. [Online]. Available at: https://arxiv.org/pdf/1710.09412.
[20] A. N. E. Xplorer and A. F. Ramework, “iGraphMix: Input Graph Mixup Method For Node Classification,” arXiv, pp. 1–11, 2023, [Online]. Available at: https://openreview.net/forum?id=a2ljjXeDcE¬eId=a2ljjXeDcE.
[21] Z. Liu, S. Li, G. Wang, C. Tan, L. Wu, and S. Z. Li, “Harnessing Hard Mixed Samples with Decoupled Regularizer,” in Advances in Neural Information Processing Systems, 2022, no. DM, pp. 1–23, [Online]. Available at: http://arxiv.org/abs/2203.10761.
[22] M. Kang, M. Kang, and S. Kim, “Catch-Up Mix: Catch-Up Class for Struggling Filters in CNN,” Proc. AAAI Conf. Artif. Intell., vol. 38, no. 3, pp. 2705–2713, Mar. 2024, doi: 10.1609/aaai.v38i3.28049.
[23] M. Kang and S. Kim, “GuidedMixup: An Efficient Mixup Strategy Guided by Saliency Maps,” Proc. AAAI Conf. Artif. Intell., vol. 37, no. 1, pp. 1096–1104, Jun. 2023, doi: 10.1609/aaai.v37i1.25191.
[24] S. Hong, Y. Yoon, H. Joo, and J. Lee, “GBMix: Enhancing Fairness by Group-Balanced Mixup,” IEEE Access, vol. 12, pp. 18877–18887, 2024, doi: 10.1109/ACCESS.2024.3358275.
[25] J. Maurya, K. R. Ranipa, O. Yamaguchi, T. Shibata, and D. Kobayashi, “Domain Adaptation using Self-Training with Mixup for One-Stage Object Detection,” in 2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Jan. 2023, pp. 4178–4187, doi: 10.1109/WACV56688.2023.00417.
[26] D. Ryou, I. Ha, H. Yoo, D. Kim, and B. Han, “Robust Image Denoising Through Adversarial Frequency Mixup,” in 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2024, pp. 2723–2732, doi: 10.1109/CVPR52733.2024.00263.
[27] K. Aggarwal and J. Srivastava, “Embarrassingly Simple Mixup for Time-series,” arXiv, pp. 1–11, Apr. 2023. [Online]. Available at: https://arxiv.org/pdf/2304.04271.
[28] P. Guo, H. Yang, and A. Sano, “Empirical Study of Mix-based Data Augmentation Methods in Physiological Time Series Data,” in 2023 IEEE 11th International Conference on Healthcare Informatics (ICHI), Jun. 2023, pp. 206–213, doi: 10.1109/ICHI57859.2023.00037.
[29] L. N. Darlow, A. Joosen, M. Asenov, Q. Deng, J. Wang, and A. Barker, “TSMix: time series data augmentation by mixing sources,” in Proceedings of the 3rd Workshop on Machine Learning and Systems, May 2023, pp. 109–114, doi: 10.1145/3578356.3592584.
[30] H. Zhou et al., “Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting,” Proc. AAAI Conf. Artif. Intell., vol. 35, no. 12, pp. 11106–11115, May 2021, doi: 10.1609/aaai.v35i12.17325.
[31] A. Bagnall et al., “The UEA multivariate time series classification archive, 2018,” arXiv, pp. 1–36, Oct. 2018. [Online]. Available at: https://arxiv.org/pdf/1811.00075.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
___________________________________________________________
International Journal of Advances in Intelligent Informatics
ISSN 2442-6571 (print) | 2548-3161 (online)
Organized by UAD and ASCEE Computer Society
Published by Universitas Ahmad Dahlan
W: http://ijain.org
E: info@ijain.org (paper handling issues)
andri.pranolo.id@ieee.org (publication issues)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0