Short-term wind power prediction based on multiple kernel extreme learning machine method
Liancai Ma, Jun Li · 2020 7th International Forum on Electrical Engineering and Automation (IFEEA) · 2020
For short-term wind power prediction, a multiple kernel extreme learning machine (MKELM) is proposed based on single kernel extreme learning machine (KELM) and multiple kernel learning (MKL) method. Compared with KELM method, in view of the deficiency of the existing single KELM regression ability, several different kernel functions are weighted combined according to the actual situation, and the unique advantages of different kernel functions are fully utilized to further improve the performance of the network. In order to verify the effectiveness of the MKELM method proposed in this paper, the MKELM method is applied to the research of short-term wind power forecasting in a certain region, and is compared with the existing extreme learning machine (ELM), KELM and support vector machine (SVM) under the same conditions. The experimental results demonstrate that the proposed MKELM method is better than KELM method in prediction accuracy, which shows its effectiveness.