SEED: Spectral Entropy-Guided Evaluation of SpatialTemporal Dependencies for Multivariate Time Series Forecasting

2025-12-09

AI Summary

Existing multivariate time series forecasting methods face three key challenges: (1) strong temporal autocorrelation is vulnerable to interference from irrelevant variables; (2) softmax-based normalization fails to model—and may even distort—negative correlations; and (3) variables lack intrinsic temporal position awareness. To address these, we propose a spectral entropy-guided dynamic dependency assessment framework. Our approach introduces four novel components: a dependency estimator, a spectral entropy fusion module, a signed graph constructor, and a contextual spatial extractor. Notably, we are the first to incorporate spectral entropy for quantifying dependency strength, explicitly preserving negative correlations while adaptively balancing channel independence and interdependence. Extensive experiments on 12 real-world datasets across diverse domains demonstrate consistent superiority over state-of-the-art methods, with significant improvements in long-horizon forecasting accuracy and generalization capability.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Effective multivariate time series forecasting often benefits from accurately modeling complex inter-variable dependencies. However, existing attention- or graph-based methods face three key issues: (a) strong temporal self-dependencies are often disrupted by irrelevant variables; (b) softmax normalization ignores and reverses negative correlations; (c) variables struggle to perceive their temporal positions. To address these, we propose extbf{SEED}, a Spectral Entropy-guided Evaluation framework for spatial-temporal Dependency modeling. SEED introduces a Dependency Evaluator, a key innovation that leverages spectral entropy to dynamically provide a preliminary evaluation of the spatial and temporal dependencies of each variable, enabling the model to adaptively balance Channel Independence (CI) and Channel Dependence (CD) strategies. To account for temporal regularities originating from the influence of other variables rather than intrinsic dynamics, we propose Spectral Entropy-based Fuser to further refine the evaluated dependency weights, effectively separating this part. Moreover, to preserve negative correlations, we introduce a Signed Graph Constructor that enables signed edge weights, overcoming the limitations of softmax. Finally, to help variables perceive their temporal positions and thereby construct more comprehensive spatial features, we introduce the Context Spatial Extractor, which leverages local contextual windows to extract spatial features. Extensive experiments on 12 real-world datasets from various application domains demonstrate that SEED achieves state-of-the-art performance, validating its effectiveness and generality.

Problem

• Addresses disruption of temporal self-dependencies by irrelevant variables.
• Overcomes softmax limitations in handling negative correlations.
• Enables variables to perceive temporal positions for spatial features.

Innovation

• Spectral entropy guides dynamic dependency evaluation
• Signed graph constructor preserves negative correlations
• Context spatial extractor captures temporal-aware features