PriceFM: Foundation Model for Probabilistic Electricity Price Forecasting

Runyao Yu, Chenhui Gu, Jochen Stiasny, Qingsong Wen, Wasim Sarwar Dilov, Lianlian Qi, Jochen Lorenz Cremer

Keywords: Day-Ahead Electricity Market, Electricity Dataset, Foundation Model, Probabilistic Forecasting

TL;DR: We present a comprehensive dataset and propose PriceFM, a foundation model with generic structure for probabilistic electricity price forecasting.

Abstract: Electricity price forecasting in Europe presents unique challenges due to the continent’s increasingly integrated and physically interconnected power market. While recent advances in deep learning and foundation models have led to substantial improvements in general time series forecasting, most existing approaches fail to capture the complex spatial interdependencies and uncertainty inherent in electricity markets. In this paper, we address these limitations by introducing a comprehensive and up-to-date dataset across 24 European countries (38 regions), spanning from 2022-01-01 to 2025-01-01. Building on this groundwork, we propose PriceFM, a spatiotemporal foundation model that integrates graph-based inductive biases to capture spatial interdependencies across interconnected electricity markets. The model is designed for multi-region, multi-timestep, and multi-quantile probabilistic electricity price forecasting. Extensive experiments and ablation studies confirm the model’s effectiveness, consistently outperforming competitive baselines and highlighting the importance of spatial context in electricity markets.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 6341