Multi-Feature Short-Term Electricity Load Forecasting Based On Improved CNN-LSTM Combined Forecasting Model

In order to maintain the stable operation of the power system and effectively guarantee the security, economy and reliability of energy supply, accurate load forecasting plays a crucial role. In order to improve the accuracy of short-term power load, this paper firstly deals with the problem of multi-factor coupling affecting the input data by principal component analysis, then optimizes the combination model of convolutional neural network and long short-term memory network by adopting an improved sparrow search algorithm, and further explores the internal correlation of the data by using the attention mechanism, and then optimizes the prediction error by using the extreme gradient boosting algorithm, and finally carries out a validation experiment on Python platform. Finally, validation experiments are carried out on Python platform to verify the effectiveness and superiority of the strategy proposed in this paper for improving the accuracy of short-term power load prediction by combining the prediction effect graphs and various error data.