Wind Speed Forecasting Model Based on Fuzzy Manifold Support Vector Machine

摘要

Support Vector Machine (SVM) is widely used in wind speed forecasting. With the application gets more intensive, two shortcomings have appeared: one is it is too sensitive to noises, and the other is it can not fully use the information included in the samples. In view of this, Fuzzy Manifold Support Vector Machine (FMSVM) is proposed in this paper. In FMSVM, the fuzzy techniques are introduced to decrease the influence of noises. Meanwhile, FMSVM takes boundary data between classes, data distributions and data manifold into consideration, so it performs better than SVM through the comparative experiments on the wind dataset of a certain wind farm.