A new modeling method based on bagging ELM for day-ahead electricity price prediction

Aiming at the shortages of traditional neural networks, a new modeling method based on Bagging ELM is proposed to establish the electricity price prediction model. The characters of day-ahead electricity price are analyzed and a novel neural network algorithm ELM is selected for its better performance to establish the basic day-ahead electricity price prediction model. Motivated by the ensemble ideas, a Bagging ensemble scheme is used to combining the single ELM learning machines. And the new Bagging ELM modeling approach is used to establish the prediction model. The day-ahead electricity price prediction model is tested by the real data. The experiments demonstrate that the new prediction model established by the new Bagging ELM modelling method has better performance.