arXiv：cs.LG· Zhenghua Pan, Ahmed Aziz Ezzat·· 4 小时前AI 评分27
评估时间序列基础模型在电价预测中的表现：污染风险、分布偏移与协变量依赖
Evaluating Time Series Foundation Models for Electricity Price Forecasting: Contamination Risk, Distributional Shifts, and Covariate Dependence
AI 导读
研究提出双数据集基准框架评估时间序列基础模型（TSFMs）在电价预测中的表现，以降低数据污染风险。结果显示 TSFMs 零样本预测具竞争力且常优于通用基线，但表现高度依赖协变量支持，未能稳定超越电价预测专用方法。TSFMs 与专用方法的简单集成展现出显著潜力，表明两者可捕捉互补的预测信息。
正文
Abstract:Time series foundation models (TSFMs) have shown strong zero-shot forecasting performance, but their generalization in covariate-driven, non-stationary settings is underexplored. Electricity price forecasting (EPF) presents a challenging testbed due to complex temporal dependencies, distributional shifts, and strong reliance on structural and contextual information. We propose a two-dataset-benchmarking framework for EPF to mitigate contamination risk and enable fair evaluation of TSFMs. We examine key aspects of EPF including point and probabilistic forecasting performance, tail behavior, price spikes, and comparisons against domain-specific methods. We find that TSFMs are highly competitive and often outperform general-purpose baselines. Yet, their performance depends critically on covariate support, and they do not consistently surpass domain-specific methods tailored to EPF. Interestingly, simple ensembles of TSFMs and domain-specific methods appear to have significant potential, suggesting that the two approaches capture complementary predictive information.
| Subjects: | Machine Learning (cs.LG); Systems and Control (eess.SY) |
| Cite as: | arXiv:2607.02623 [cs.LG] |
| (or arXiv:2607.02623v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.02623 arXiv-issued DOI via DataCite |
|
| Journal reference: | ICML 2026 Foundation Models for Structured Data Workshop, 43rd International Conference on Machine Learning (ICML), Seoul, South Korea |
Submission history
From: Ahmed Aziz Ezzat [view email]
[v1]
Thu, 2 Jul 2026 11:43:22 UTC (2,185 KB)
[v2]
Wed, 7 Oct 2026 14:43:57 UTC (2,185 KB)
来源：arXiv：cs.LG · arxiv.org