Research on Electricity Market Clearing Price Prediction Based on Modal Decomposition and Machine Learning

Accurate prediction of electricity market clearing prices provides effective guidance for electricity market operation and policy formulation. This study proposes a framework for predicting electricity market clearing prices based on modal decomposition and feature fusion. Firstly, a modal decomposition method for electricity market clearing prices based on STL model is proposed. The periodic time series data of clearing prices are successfully decomposed into trend components, periodic components, and residual components. Pearson correlation analysis and feature importance analysis methods are then employed to construct a clearing price and its component prediction model. Subsequently, the ISSA Informer model, ISSA BiGRU model, and ISSA-GRU model are developed to accurately predict the trend components, periodic components, and residual components, respectively. On this basis, the STL-ISSA-ML model is constructed using feature fusion method, achieving an R²of 0.9894, an MAE of 11.33, and a MAPE of 2.92%the model for electricity market clearing price forecasting. Finally, the modified CSTL-ISSA-ML model is obtained through time series analysis and error correction methods. The enhanced model achieves an R² of 0.9986, an MAE of 4.72, and a MAPE of 1.47%. Compared with the ISSA Informer model, ISSA BiGRU model, and ISSA-GRU model, the prediction performance is optimized by an average of 59.16%, 69.18%, and 81.68%, respectively, thereby validating the effectiveness of the proposed modal decomposition and feature fusion prediction framework The research findings provides theoretical support and practical reference for the healthy development of the electricity market.