Abstract
Deep learning methods can fit the observation history over different time series with multiple levels of representations from huge dataset. However, it is challenging to directly train deep neural networks on a raw dataset with a large number of time series, as the different time-series have diverse scales. We initiate the study of an effective deep residual framework named MIR-TS for time series prediction with multi-output integration on time series data with diverse scales. Specifically, we leverage the residual module that constrains the original input average close to 0 to transform the original input, so that the distribution of features changes from sparse to dense. Compared with the traditional residual network, this approach improves the generalization of model via residual reuse, capturing more detailed features of time series to improve prediction. The results on the M3 and TOURISM benchmarks show that MIR-TS achieves a consistent better or highly comparable performance across different time series frequencies.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Assimakopoulos, V., Nikolopoulos, K.: The theta model: a decomposition approach to forecasting. Int. J. Forecast. 16(4), 521–530 (2000)
Athanasopoulos, G., Hyndman, R.J., Song, H., Wu, D.C.: The tourism forecasting competition. Int. J. Forecast. 27(3), 822–844 (2011)
Baker, L.C., Howard, J.: Winning methods for forecasting tourism time series. Int. J. Forecast. 27(3), 850–852 (2011)
Bandara, K., Hewamalage, H., Liu, Y., Kang, Y., Bergmeir, C.: Improving the accuracy of global forecasting models using time series data augmentation. Pattern Recogn. 120, 108148 (2021)
Benidis, K., et al.: Neural forecasting: introduction and literature overview. CoRR abs/2004.10240 (2020)
Box, G.E.P., Jenkins, G.M., MacGregor, J.F.: Some recent advances in forecasting and control. J. Roy. Stat. Soc. Ser. C (Appl. Stat.) 23(2), 158–179 (1974)
Breiman, L.: Bagging predictors. Mach. Learn. 24(2), 123–140 (1996)
Brierley, P.: Winning methods for forecasting seasonal tourism time series. Int. J. Forecast. 27(3), 853–854 (2011)
Fiorucci, J.A., Pellegrini, T.R., Louzada, F., Petropoulos, F., Koehler, A.B.: Models for optimising the theta method and their relationship to state space models. Int. J. Forecast. 32(4), 1151–1161 (2016)
Flunkert, D.S.V., Gasthaus, J., Januschowski, T.: DeepAR: probabilistic forecasting with autoregressive recurrent networks. Int. J. Forecast. 36(3), 1181–1191 (2020)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, pp. 770–778 (2016)
Holt, C.C.: Forecasting seasonals and trends by exponentially weighted moving averages. Int. J. Forecast. 20(1), 5–10 (2004)
Hu, H., Tang, M., Bai, C.: DATSING: data augmented time series forecasting with adversarial domain adaptation. In: CIKM 2020: The 29th ACM International Conference on Information and Knowledge Management, pp. 2061–2064 (2020)
Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, pp. 2261–2269 (2017)
Li, S., et al.: Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. In: Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, pp. 5244–5254 (2019)
van der Maaten, L., Hinton, G.: Visualizing data using t-SNE. J. Mach. Learn. Res. 9(86), 2579–2605 (2008)
Makridakis, S., Hibon, M.: The M3-competition: results, conclusions and implications. Int. J. Forecast. 16(4), 451–476 (2000)
Makridakis, S., Spiliotis, E., Assimakopoulos, V.: The M4 competition: results, findings, conclusion and way forward. Int. J. Forecast. 34(4), 802–8081 (2018)
Makridakis, S., Spiliotis, E., Assimakopoulos, V.: The M4 competition: 100,000 time series and 61 forecasting methods. Int. J. Forecast. 36(1), 54–74 (2020)
Montori, F., Liao, K., Jayaraman, P.P., Bononi, L., Sellis, T., Georgakopoulos, D.: Classification and annotation of open internet of things datastreams. In: Hacid, H., Cellary, W., Wang, H., Paik, H.-Y., Zhou, R. (eds.) WISE 2018, Part II. LNCS, vol. 11234, pp. 209–224. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-02925-8_15
van den Oord, A., et al.: WaveNet: a generative model for raw audio. In: The 9th ISCA Speech Synthesis Workshop, p. 125 (2016)
Oreshkin, B.N., Carpov, D., Chapados, N., Bengio, Y.: N-BEATS: neural basis expansion analysis for interpretable time series forecasting. In: 8th International Conference on Learning Representations, ICLR 2020 (2020)
Smyl, S.: A hybrid method of exponential smoothing and recurrent neural networks for time series forecasting. Int. J. Forecast. 36(1), 75–85 (2020)
Smyl, S., Kuber, K.: Data preprocessing and augmentation for multiple short time series forecasting with recurrent neural networks. In: 36th International Symposium on Forecasting (2016)
Spiliotis, E., Assimakopoulos, V., Nikolopoulos, K.: Forecasting with a hybrid method utilizing data smoothing, a variation of the theta method and shrinkage of seasonal factors. Int. J. Prod. Econ. 209, 92–102 (2019)
Sun, H., et al.: Fast anomaly detection in multiple multi-dimensional data streams. In: 2019 IEEE International Conference on Big Data (IEEE BigData), pp. 1218–1223 (2019)
Wang, L., Wang, Z., Qu, H., Liu, S.: Optimal forecast combination based on neural networks for time series forecasting. Appl. Soft Comput. 66, 1–17 (2018)
Wang, Y., Smola, A., Maddix, D.C., Gasthaus, J., Foster, D., Januschowski, T.: Deep factors for forecasting. In: Proceedings of the 36th International Conference on Machine Learning, ICML 2019, pp. 6607–6617 (2019)
Winters, P.R.: Forecasting sales by exponentially weighted moving averages. Manag. Sci. 6(3), 324–342 (1960)
Acknowledgements
This work is supported by the National Natural Science Foundation of China (No. 61832001) and Australian Research Council (No. DP220101420).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Li, H., Tang, M., Liao, K., Shao, J. (2022). A Multi-output Integration Residual Network for Predicting Time Series Data with Diverse Scales. In: Khanna, S., Cao, J., Bai, Q., Xu, G. (eds) PRICAI 2022: Trends in Artificial Intelligence. PRICAI 2022. Lecture Notes in Computer Science, vol 13629. Springer, Cham. https://doi.org/10.1007/978-3-031-20862-1_28
Download citation
DOI: https://doi.org/10.1007/978-3-031-20862-1_28
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-20861-4
Online ISBN: 978-3-031-20862-1
eBook Packages: Computer ScienceComputer Science (R0)