Research on Short-term Power Generation Forecasting Based on Ensemble Learning

Accurate forecasting of short-term power generation is the key to improve the efficiency of electric power company. To address the issue of inadequate prediction accuracy by current models, a new short-term power generation forecasting model based on ensemble learning (MSTLEnsemble) was proposed in this paper by combining the complementary advantages of seasonal trend decomposition of local weighted regression, variational mode decomposition, short and long time memory neural network and multiple linear regression. And the model's effectiveness is verified through comparative experiments.