Electricity Price Prediction Based on LSTM and LightGBM

Abstract

In the open electricity market, short-term electricity price forecasting is a significant research direction. At present, a single prediction model will have different prediction deviations when predicting. This article proposes a method to simultaneously input the original data into the LSTM network and the LightGBM model. Simultaneously. Models with higher prediction limits. Experiments have proved that the combined model can effectively increase the lower limit of prediction.