Short-term Wind Power Prediction Based on Soft Margin Multiple Kernel Learning Method

For short-term wind power prediction, a soft margin multiple kernel learning (MKL) method is proposed. In order to improve the predictive effect of MKL for power, slack variable introduced into each base solve objective function. Two kinds methods based on hinge loss function and square can be obtained when functions are selected. The improved demonstrate good robustness avoid disadvantage hard which only selects few kernels discards other useful solving function, thereby achieving an effective yet sparse solution method. verify effectiveness proposed method, was applied second farm Tianfeng from Xinjiang single-step multi-step predictions also carried out using measured data provided by alberta electric system operator (AESO). Compared with support vector machine (SVM), extreme (ELM), (KELM) as well SimpleMKL under same conditions, experimental results that different efficiently achieve higher prediction accuracy generalization performance confirms