Research on Short term Load Forecasting Based on CNN-LSTM-Attention
Cheng Xie, Pengfei Zhang, Jiarong Tao · 2025
With the increasing demand for electricity and the rising requirements for prediction accuracy, short-term power system load forecasting is crucial for ensuring the safe and stable operation of power systems as well as enhancing their economic efficiency. However, traditional methods suffer from redundant information. Therefore, this paper proposes a combined network prediction method based on CNN-LSTM-Attention, which utilizes LSTM networks for load forecasting and incorporates CNN and attention mechanisms to optimize prediction accuracy. Experimental results demonstrate that the combined network outperforms other comparative prediction models in terms of prediction accuracy across different seasons.