Scholarlink

<p>In the context of electricity market reform, precise short-term electricity price prediction is essential for optimizing market trading mechanisms and enhancing decision-making processes. This paper presents a novel prediction method that integrates MIC (maximum information coefficient) analysis with a TCN-CBAM (improved temporal convolution network). Initially, MIC correlation analysis is employed to identify the relationships between electricity price series and influencing factors. Subsequently, fuzzy C-means clustering is applied to group the electricity price series, providing a solid foundation for further analysis of price fluctuations. To enhance prediction accuracy, VMD (variational mode decomposition) is utilized to decompose the highly volatile electricity price data into smoother modal components. Finally, the TCN-CBAM model, incorporating both channel and spatial attention mechanisms, is applied to predict the components. The experimental results demonstrate a significant improvement in prediction accuracy over traditional methods, affirming the efficacy of this approach in short-term electricity price forecasting. This method offers a robust tool for supporting market decision-making and optimizing electricity price predictions in power markets.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Short-term Electricity Price Prediction Based on Maximal Information Coefficient and Improved FCM-VMD-TCN

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月5日

查看全文
  • Lantao Jing,
  • Xu Li,
  • WeiDong Chen,
  • Ye Zhang,
  • Liang Wang,
  • Rui Tian,
  • Yi Zhao

摘要

In the context of electricity market reform, precise short-term electricity price prediction is essential for optimizing market trading mechanisms and enhancing decision-making processes. This paper presents a novel prediction method that integrates MIC (maximum information coefficient) analysis with a TCN-CBAM (improved temporal convolution network). Initially, MIC correlation analysis is employed to identify the relationships between electricity price series and influencing factors. Subsequently, fuzzy C-means clustering is applied to group the electricity price series, providing a solid foundation for further analysis of price fluctuations. To enhance prediction accuracy, VMD (variational mode decomposition) is utilized to decompose the highly volatile electricity price data into smoother modal components. Finally, the TCN-CBAM model, incorporating both channel and spatial attention mechanisms, is applied to predict the components. The experimental results demonstrate a significant improvement in prediction accuracy over traditional methods, affirming the efficacy of this approach in short-term electricity price forecasting. This method offers a robust tool for supporting market decision-making and optimizing electricity price predictions in power markets.