Abstract
Improving long-term time series forecasting accuracy and efficiency is of great value for real-world applications. The main challenge in the long-term forecasting of multivariate time series is to accurately capture the local dynamics and long-term dependencies of time series. Currently, most approaches capture temporal dependencies and inter-variable dependencies in intertwined temporal patterns, which are unreliable. Moreover, models based on time series decomposition methods are still unable to capture both short- and long-term dependencies well. In this paper, we propose an efficient multivariate time series forecasting model CNformer with three distinctive features. (1) The CNformer is a fully CNN-based time series forecasting model. (2) In the encoder, the stacked dilated convolution as a built-in block is combined with the time series decomposition to extract the seasonal component of the time series. (3) The convolution-based encoder-decoder attention mechanism refines seasonal patterns in the decoder and captures complex combinations between different related time series. Owing to these features, our CNformer has lower memory and time overhead than models based on self-attention and the Auto-Correlation mechanism. Experimental results show that our model achieves state-of-the-art performance on four real-world datasets, with a relative performance improvement of 20.29%.
Similar content being viewed by others
References
Yamak PT, Yujian L, Gadosey PK (2019) A comparison between arima, lstm, and gru for time series forecasting. In: Proceedings of the 2019 2nd international conference on algorithms, computing and artificial intelligence, pp 49–55
Smyl S (2020) A hybrid method of exponential smoothing and recurrent neural networks for time series forecasting. Int J Forecast 36(1):75–85
Song W, Fujimura S (2021) Capturing combination patterns of long-and short-term dependencies in multivariate time series forecasting. Neurocomputing 464:72–82
Yin C, Dai Q (2022) A deep multivariate time series multistep forecasting network. Appl Intell 52(8):8956–8974
Lai G, Chang W-C, Yang Y, Liu H (2018) Modeling long-and short-term temporal patterns with deep neural networks. In: The 41st International ACM SIGIR conference on research & development in information retrieval, pp 95–104
Salinas D, Flunkert V, Gasthaus J, Januschowski T (2020) Deepar: probabilistic forecasting with autoregressive recurrent networks. Int J Forecast 36(3):1181–1191
Niu H, Xu K, Wang W (2020) A hybrid stock price index forecasting model based on variational mode decomposition and lstm network. Appl Intell 50(12):4296–4309
Chang Y-Y, Sun F-Y, Wu Y-H, Lin S-D A memory-network based solution for multivariate time-series forecasting, arXiv:1809.02105
Qin Y, Song D, Cheng H, Cheng W, Jiang G, Cottrell GW (2017) A dual-stage attention-based recurrent neural network for time series prediction. In: Proceedings of the 26th international joint conference on artificial intelligence, pp 2627–2633
Shih S-Y, Sun F-K, Lee H-y (2019) Temporal pattern attention for multivariate time series forecasting. Mach Learn 108(8):1421–1441
Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser Ł, Polosukhin I (2017) Attention is all you need. In: Advances in neural information processing systems, pp 5998–6008
Wang W, Xie E, Li X, Fan D-P, Song K, Liang D, Lu T, Luo P, Shao L (2021) Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 568–578
Chu X, Tian Z, Wang Y, Zhang B, Ren H, Wei X, Xia H, Shen C (2021) Twins: Revisiting the design of spatial attention in vision transformers. Adv Neural Inf Process Syst 34:9355–9366
Zhou H, Zhang S, Peng J, Zhang S, Li J, Xiong H, Zhang W (2021) Informer: Beyond efficient transformer for long sequence time-series forecasting. In: Proceedings of AAAI
Li S, Jin X, Xuan Y, Zhou X, Chen W, Wang Y-X, Yan X (2019) Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. Adv Neural Inf Process Syst 32:5243–5253
Kitaev N, Kaiser L, Levskaya A (2019) Reformer: the efficient transformer. In: International conference on learning representations
Bandara K, Bergmeir C, Hewamalage H (2020) Lstm-msnet: Leveraging forecasts on sets of related time series with multiple seasonal patterns. IEEE Trans Neural Netw Learn Syst 32(4):1586–1599
Wu H, Xu J, Wang J, Long M (2021) Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting. Adv Neural Inf Process Syst 34:22419– 22430
He X, Shi S, Geng X, Xu L (2022) Information-aware attention dynamic synergetic network for multivariate time series long-term forecasting. Neurocomputing 500:143–154
Bi H, Lu L, Meng Y (2022) Hierarchical attention network for multivariate time series long-term forecasting. Appl Intell :1–12
Karanikola A, Liapis CM, Kotsiantis S (2022) A comparison of contemporary methods on univariate time series forecasting. In: Advances in machine learning/deep learning-based technologies, Springer, pp 143–168
Hajmohammadi H, Heydecker B (2021) Multivariate time series modelling for urban air quality. Urban Clim 37:100834
Fathian F, Fard AF, Ouarda TB, Dinpashoh Y, Nadoushani SM (2019) Modeling streamflow time series using nonlinear setar-garch models. J Hydrol 573:82–97
Zhang M, Jiang X, Fang Z, Zeng Y, Xu K (2019) High-order hidden markov model for trend prediction in financial time series. Phys A Stat Mech Appl 517:1–12
Rangapuram SS, Seeger M, Gasthaus J, Stella L, Wang Y, Januschowski T (2018) Deep state space models for time series forecasting. In: Proceedings of the 32nd international conference on neural information processing systems, pp 7796–7805
Martínez F, Charte F, Frías MP, Martínez-Rodríguez AM (2022) Strategies for time series forecasting with generalized regression neural networks. Neurocomputing 491:509–521
Chen W, Xu H, Chen Z, Jiang M (2021) A novel method for time series prediction based on error decomposition and nonlinear combination of forecasters. Neurocomputing 426:85–103
Chen Z, Ma Q, Lin Z (2021) Time-aware multi-scale rnns for time series modeling. In: IJCAI
Yang T, Yu X, Ma N, Zhao Y, Li H (2021) A novel domain adaptive deep recurrent network for multivariate time series prediction. Eng Appl Artif Intell 106:104498
Yang Y, Fan C, Xiong H (2022) A novel general-purpose hybrid model for time series forecasting. Appl Intell 52(2):2212–2223
Huang S, Wang D, Wu X, Tang A (2019) Dsanet: Dual self-attention network for multivariate time series forecasting. In: Proceedings of the 28th ACM international conference on information and knowledge management, pp 2129–2132
Yazici I, Beyca OF, Delen D (2022) Deep-learning-based short-term electricity load forecasting: A real case application. Eng Appl Artif Intell 109:104645
Bai S, Kolter JZ, Koltun V An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv:1803.01271
Fan J, Zhang K, Huang Y, Zhu Y, Chen B (2021) Parallel spatio-temporal attention-based TCN for multivariate time series prediction. Neural Comput Applic :1–10
Kenton JDM-WC, Toutanova LK (2019) Bert: Pre-training of deep bidirectional transformers for language understanding. In: Proceedings of NAACL-HLT, pp 4171–4186
Liu Z, Lin Y, Cao Y, Hu H, Wei Y, Zhang Z, Lin S, Guo B (2021) Swin transformer: Hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 10012–10022
Shen L, Wang Y (2022) Tcct: Tightly-coupled convolutional transformer on time series forecasting. Neurocomputing 480:131–145
Lara-benítez P, Carranza-García M, Luna-Romera JM, Riquelme JC (2020) Temporal convolutional networks applied to energy-related time series forecasting. Appl Sci 10(7):2322
Fang X, Yuan Z (2019) Performance enhancing techniques for deep learning models in time series forecasting. Eng Appl Artif Intell 85:533–542
Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, Killeen T, Lin Z, Gimelshein N, Antiga L et al (2019) Pytorch: an imperative style, high-performance deep learning library. In: Proceedings of the 33rd international conference on neural information processing systems, pp 8026–8037
Kingma DP, Ba J (2015) Adam: a method for stochastic optimization. In: International conference on learning representations
Funding
This work is partially supported by a grant from the National Natural Science Foundation of China (No. 62272368, No. 62032017), the Innovation Capability Support Program of Shaanxi (No. 2023-ZC-TD-0008), the Key Research and Development Program of Shaanxi (No. 2021ZDLGY03-09, No. 2021ZDLGY07-02, No. 2021ZDLGY07-03), Shaanxi Qinchuangyuan “scientists+engineers” team in 2023 (No. 41), and The Youth Innovation Team of Shaanxi Universities.
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Conflict of Interests
The authors have no competing financial interests to declare that are relevant to the content of this article.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Wang, X., Liu, H., Yang, Z. et al. CNformer: a convolutional transformer with decomposition for long-term multivariate time series forecasting. Appl Intell 53, 20191–20205 (2023). https://doi.org/10.1007/s10489-023-04496-6
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-023-04496-6