Nested Spatio-Temporal Time Series Forecasting

TL;DR

This paper introduces a nested spatiotemporal forecasting framework that combines macro-level regional trends with micro-level observations, improving accuracy in noisy, non-stationary environments.

Contribution

It proposes a spectral clustering-based regional representation and a progressive coarse-to-fine predictor for enhanced spatiotemporal forecasting.

Findings

Outperforms state-of-the-art baselines on multiple datasets.

Effectively filters systematic noise while preserving essential trends.

Anticipates dynamic anomalies like periodic offsets in advance.

Abstract

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.