Short-term Electricity Price Forecasting Method Based on Quadratic Decomposition and ATT-CNN-LSTM-MLR Combined Model
Targeting the problem of the strong volatility of electricity spot market prices and the existence of jump points and peak points in electricity prices,the paper proposes a short-term electricity price prediction model based on a combination of quadratic decomposition and attention mechanism(ATT)-convolutional neural network(CNN)-long short-term memory neural network(LSTM)-multiple linear regression(MLR).Firstly,the adaptive noise complete set empirical mode decomposition and variational mode decomposition are used to decompose the original electricity price sequence into a series of modal components,exploring the features of the jump points and peak points in the electricity prices.Secondly,for the decomposed low-frequency and high-frequency subsequences,MLR shallow learning model and ATT-CNN-LSTM combined model are respectively established to achieve multi frequency prediction,and the output of each sub model is reconstructed to obtain the final prediction results.Finally,a case study is conducted using data with strong volatility characteristics from European Electricity Exchange to validate the effectiveness of the model.
万方数据
维普
