AUTHOR=Lin Qingbiao , Chen Wan , Zhao Xu , Zhou Shangchou , Gong Xueliang , Zhao Bo 

TITLE=Research on a price prediction model for a multi-layer spot electricity market based on an intelligent learning algorithm

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 12 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2024.1308806

DOI=10.3389/fenrg.2024.1308806

ISSN=2296-598X

ABSTRACT=As the unified electricity market in the southern region advances, the demand for accurate electricity spot market price prediction intensifies among market participants. This paper addresses this imperative by presenting a sophisticated multi-level spot electricity price prediction model. Grounded in intelligent learning algorithms and aligned with market principles, our methodology begins by identifying influential factors shaping spot market prices. A model is then introduced to identify historically analogous days, aiding the selection of pertinent historical market information. The CEEMDAN-XGT model is employed to enhance predictive accuracy by decomposing historical data, mitigating volatility-induced forecasting errors. The marine hunting algorithm optimizes the CNN-BiLSTM model, resulting in superior performance metrics. Comparative analysis demonstrates the model's efficacy with an SSE of 0.5828, MAE of 0.0474, and an R-squared value of 0.9466. This superior performance underscores the practical utility of the proposed model, positioning it as a valuable auxiliary information processing tool for market participants in formulating informed market decisions.