Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Nested Spatio-Temporal Time Series Forecasting

About

Spatiotemporal forecasting is critical for real-world applications like traffic management, yet capturing reliable interactions remains challenging under noisy and non-stationary conditions. Existing methods primarily rely on historical spatial priors, often failing to account for evolving temporal correlations and suffering from systematic errors. In this work, we propose a nested forecasting framework that couples future macro-level regional trends with micro-level historical observations, enabling top-down guidance from abstract future representations for fine-grained forecasting. Specifically, we employ a spectral clustering-based approach to construct semantically coherent regions, providing both theoretical and empirical evidence that this representation effectively filters systematic noise while preserving essential trends. Building on this, we develop a progressive coarse-to-fine predictor to integrate these representative features into the inference process. This enables the model to leverage trend predictions to anticipate dynamic anomalies, such as periodic offsets, in advance. Furthermore, extensive experiments on multiple high-dimensional datasets demonstrate that our method consistently outperforms state-of-the-art baselines, validating the effectiveness of future macro-guided nested forecasting.

Yinghao Ai, Yukai Zhou, Ruoxi Jiang, Junyi An, Chao Qu, Zhijian Zhou, Shiyu Wang, Fenglei Cao, Zenglin Xu, Furao Shen, Yuan Qi• 2026

Related benchmarks

TaskDatasetResultRank
Long-term time-series forecastingSolar Energy
MSE0.201
126
Spatio-temporal traffic forecastingUrbanEV
RMSE10.26
75
Air quality forecastingKnowAir
MAE14.87
49
Long-term time-series forecastingElectricity
MSE0.141
22
Spatio-Temporal Time Series ForecastingGBA LargeST (test)
MAE (Horizon 3)16.05
12
Spatio-Temporal Time Series ForecastingGLA LargeST (test)
MAE (H=3)15.12
10
Spatio-Temporal Time Series ForecastingGLA
MAE23.63
10
Spatio-Temporal Time Series ForecastingCA
MAE22.05
10
Spatio-Temporal Time Series ForecastingLargeST CA (test)
MAE (Horizon 3)13.98
8
Showing 9 of 9 rows

Other info

Follow for update