Research on Electricity Price Prediction Technology Based on Mode Decomposition and Deep Learning

Electricity price forecasting is crucial in energy markets and grid management, where accurate electricity price forecasting can optimize energy production and allocation, guide market transactions, and reduce operating costs. However, the volatility and complexity of electricity prices make it difficult for traditional statistical methods and some machine learning methods to achieve high-precision prediction. To address this challenge, this paper proposes an innovative electricity price forecasting model called the VMD-Transformer model. This model combines variational mode decomposition (VMD) and a Transformer-based deep learning architecture, which decomposes complex electricity price time series data into multiple intrinsic mode functions (IMFs) through VMD, thereby extracting multi-scale features. Subsequently, using a Transformer-based deep learning model, long-distance dependencies across modes and complex temporal patterns are effectively captured through the multi-head self-attention mechanism. Experimental results show that the VMD-Transformer model outperforms existing models on multiple electricity price datasets. Further analysis and case studies demonstrate that the model exhibits higher prediction accuracy and stability during periods of severe electricity price fluctuations. The VMD-Transformer model provides an effective new method for electricity price forecasting, significantly improving prediction performance and offering strong support for energy market participants and grid operators.