UltraSTF: Ultra-Compact Model for Large-Scale Spatio-Temporal Forecasting
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911094117761024 |
|---|---|
| author | Yeh, Chin-Chia Michael Fan, Xiran Jiang, Zhimeng Fan, Yujie Chen, Huiyuan Saini, Uday Singh Lai, Vivian Dai, Xin Wang, Junpeng Zhuang, Zhongfang Wang, Liang Zheng, Yan |
| author_facet | Yeh, Chin-Chia Michael Fan, Xiran Jiang, Zhimeng Fan, Yujie Chen, Huiyuan Saini, Uday Singh Lai, Vivian Dai, Xin Wang, Junpeng Zhuang, Zhongfang Wang, Liang Zheng, Yan |
| contents | Spatio-temporal data, prevalent in real-world applications such as traffic monitoring, financial transactions, and ride-share demands, represents a specialized case of multivariate time series characterized by high dimensionality. This high dimensionality necessitates computationally efficient models and benefits from applying univariate forecasting approaches through channel-independent strategies. SparseTSF, a recently proposed competitive univariate forecasting model, leverages periodicity to achieve compactness by focusing on cross-period dynamics, extending the Pareto frontier in terms of model size and predictive performance. However, it underperforms on spatio-temporal data due to limited capture of intra-period temporal dependencies. To address this limitation, we propose UltraSTF, which integrates a cross-period forecasting component with an ultra-compact shape bank component. Our model efficiently captures recurring patterns in time series using the attention mechanism of the shape bank component, significantly enhancing its capability to learn intra-period dynamics. UltraSTF achieves state-of-the-art performance on the LargeST benchmark while utilizing fewer than 0.2% of the parameters required by the second-best methods, thereby further extending the Pareto frontier of existing approaches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_20634 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | UltraSTF: Ultra-Compact Model for Large-Scale Spatio-Temporal Forecasting Yeh, Chin-Chia Michael Fan, Xiran Jiang, Zhimeng Fan, Yujie Chen, Huiyuan Saini, Uday Singh Lai, Vivian Dai, Xin Wang, Junpeng Zhuang, Zhongfang Wang, Liang Zheng, Yan Machine Learning Artificial Intelligence Spatio-temporal data, prevalent in real-world applications such as traffic monitoring, financial transactions, and ride-share demands, represents a specialized case of multivariate time series characterized by high dimensionality. This high dimensionality necessitates computationally efficient models and benefits from applying univariate forecasting approaches through channel-independent strategies. SparseTSF, a recently proposed competitive univariate forecasting model, leverages periodicity to achieve compactness by focusing on cross-period dynamics, extending the Pareto frontier in terms of model size and predictive performance. However, it underperforms on spatio-temporal data due to limited capture of intra-period temporal dependencies. To address this limitation, we propose UltraSTF, which integrates a cross-period forecasting component with an ultra-compact shape bank component. Our model efficiently captures recurring patterns in time series using the attention mechanism of the shape bank component, significantly enhancing its capability to learn intra-period dynamics. UltraSTF achieves state-of-the-art performance on the LargeST benchmark while utilizing fewer than 0.2% of the parameters required by the second-best methods, thereby further extending the Pareto frontier of existing approaches. |
| title | UltraSTF: Ultra-Compact Model for Large-Scale Spatio-Temporal Forecasting |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2502.20634 |