Research on long-term time series prediction has primarily relied on Transformer and MLP models, while the potential of convolutional networks in this domain remains underexplored. To address this, we propose a novel multi-scale time series reshape module that effectively captures cross-period patch interactions and variable dependencies. Building on this, we develop MS-DFTVNet, the multi-scale 3D deformable convolutional framework tailored for long-term forecasting. Moreover, to handle the inherently uneven distribution of temporal features, we introduce a context-aware dynamic deformable convolution mechanism, which further enhances the model's ability to capture complex temporal patterns. Extensive experiments demonstrate that MS-DFTVNet not only significantly outperforms strong baselines but also achieves an average improvement of about 7.5% across six public datasets, setting new state-of-the-art results.
@article{arxiv.2506.17253,
title = {MS-DFTVNet:A Long-Term Time Series Prediction Method Based on Multi-Scale Deformable Convolution},
author = {Chenghan Li and Mingchen Li and Yipu Liao and Ruisheng Diao},
journal= {arXiv preprint arXiv:2506.17253},
year = {2025}
}