Ensemble forecasting for electricity consumption based on nonlinear optimization
Jun Hao, Qianqian Feng, Weilan Suo, Guowei Gao, Xiaolei L. Sun · Procedia Computer Science · 2019
Accurate electricity power demand forecasting can provide scientific decision-making basis for policy making and planning implementation and the electricity-generating target. In this paper, a novel ensemble forecasting model with nonlinear optimization is proposed to predict the demand of electricity. The results of basic forecasting models including exponential smoothing, ARIMA, SVR and extreme learning machine are integrated. Taking clean electricity demand of world’s major regions as sample, the results reveal that the ensemble approach performs much better than the single and average integrated models in terms of the accuracy.