Research on the Application of Gradient Lifting Stochastic Forest Model in Electricity Price Forecasting

Power generation enterprises need to predict the future electricity price in order to guide their bidding. To solve the problem of electricity price forecasting, a new gradient lifting stochastic forest model is proposed and applied to electricity price forecasting. This method is an integrated learning method. Based on the random forest model, gradient lifting algorithm is applied to effectively combine the advantages of Bagging and Boosting integrated learning strategies to improve the accuracy of model prediction. Applying this model to the electricity price forecast of PJM day-ahead market, the result shows that this model is superior to the traditional random forest model and lifting tree model, and can further improve the accuracy of electricity price forecast on its basis.