Ultra-short-term Wind Power Forecast Using Ensemble Learning and Elephant Herd Optimization Algorithm

Abstract

Accurate prediction of wind power is essential for efficient use of energy. In this paper, an ensemble learning model of optimization algorithm is proposed. Firstly, the data of wind power are decomposed into a series of signal sets by Ensemble empirical mode decomposition. Then, the least squares support vector machine (LSSVM) optimized by Elephant Herd optimization algorithm (EHO) is used to predict each component signal. Clustering method is utilized to cluster the samples. Finally, the EHO-LSSVM method is used to ensemble the sample results to get the final prediction value. Wind power data of PJM west area are used to study the effects of the hybrid method. The comparison results with eight benchmark models shows that the hybrid model has better performance and smaller error values than all other benchmark models. In conclusion, the proposed ensemble learning model is considerably effective and contains high robustness for the wind power data forecast.