Short-Term Power Load Forecasting Based on Self-Attention-LSTM

In order to improve the accuracy of power load forecasting and ensure the stable operation of the power grid. This article proposes an improved Self-Attention-LSTM combination model for power load forecasting. Firstly, analyze the nonlinear characteristics that affect power load data, and then build a Self-Attention-LSTM combination model to enhance key time step feature extraction by introducing self attention mechanism. Then, optimize the hyperparameters of the model. Finally, the prediction results of the Self-Attention-LSTM model, RNN model, and LSTM model will be compared and analyzed. The results indicate that the Self-Attention-LSTM combination model has the best prediction performance, as it can more accurately capture data features and improve the accuracy of power load forecasting.