FeTS: A Feature-Aware Framework for Time Series Forecasting

Association for the Advancement of Artificial Intelligence（AAAI）

Le Wang,  Jianyong Chen^*,  Songbai Liu

School of Computer Science and Software Engineering, Shenzhen University

Abstract

Time series forecasting faces a fundamental challenge: the uneven distribution of predictive importance in time series data, where some specific time points and feature combinations carry disproportionately predictive power. As a result, uniform processing methods that treat all data alike inevitably fall short of optimal performance. To address this problem, we propose FeTS, a feature-aware framework that comprehensively learns temporal features through two key components: (i) Adaptive Feature Extraction (AdaFE), which dynamically discovers the most important features within each temporal patch and extracts them on the fly, yielding sharper and more focused local representations; and (ii) Dual-Scale Feed-Forward Network (DSFFN), which strategically integrates fine-grained local features with global long-term dependencies to achieve richer dual-scale representation learning. Extensive experiments on eight benchmark datasets demonstrate that FeTS achieves state-of-the-art performance in time series forecasting tasks, offering a novel solution to the challenge of uneven predictive importance in forecasting.

Figure 1: Mutual Information in Patch

Figure 2: Overview of our framework.

Table 1:Multivariate long-term time series forecasting results. The best results are highlighted in bold and the second best are underlined.

Figure 3: Heatmap comparison of data at input and output of AdaFE.

Figure 4: Performance of FeTS and other models with different lookback windows. The forecast horizon is set as 96.

Figure 5:Model efficiency comparison. The running efficiency of seven models on the Weather dataset with the prediction length H = 96.

Acknowledgement

This work is partially supported by the National Natural Science Foundation of China (NSFC) under Grant 62306180, the Natural Science Foundation of Guangdong Province under Grant 2023A1515011238, the Shenzhen Science and Technology Program under Grant JCYJ20250604181503004.

Bibtex

@inproceedings{wang2026fets,

  title={FeTS: A Feature-Aware Framework for Time Series Forecasting},

  author={Wang, Le and Chen, Jianyong and Liu, Songbai},

  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},

  volume={40},

  number={31},

  pages={26328--26336},

  year={2026}

}

Downloads