.png){kind=icon}
.png){kind=icon}
Return
An accurate data-driven multi-horizon short-term electric load prediction using HODMD
.png){kind=icon}
PRE
AI
DOI：10.1007/s12065-026-01181-8.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accurate short-term electric load forecasting (STLF) is critical for reliable and economical power system operation. This paper presents a deterministic, data-driven multi-horizon STLF framework based on Higher-Order Dynamic Mode Decomposition (HODMD), which extends classical DMD through delay embedding to capture higher-order temporal dependencies in load dynamics. Unlike deep learning models, the proposed approach does not rely on iterative training or hyperparameter tuning, ensuring interpretability and computational efficiency. Extensive experiments conducted across multiple geographical regions, seasonal conditions, and forecasting horizons demonstrate that the proposed approach achieves the lowest prediction error with 10%-40% improvement compared to various statistical, machine learning, deep learning and ensemble methods. Statistical significance is validated using Friedman and Nemenyi post-hoc tests, confirming the robustness of the observed improvements. While the framework does not provide probabilistic uncertainty estimates, it offers a stable and transparent alternative for real-time multi-horizon load forecasting applications.Graphic AbstractOverview of the proposed HODMD-based multi-horizon short-term load forecasting framework.
Keywords:
Short-term load forecasting
Higher order dynamic mode decomposition
Interpretability
Hyperparameter tuning
Higher-order temporal dependencies
Journal
IF:
2.6
Papers:
116
Citations:
2.0K