Hybrid load forecasting method based on fuzzy support vector machine and linear extrapolation

For the load affected by many factors and near the far smaller feature, a hybrid load forecasting method based on fuzzy support vector machine and linear extrapolation is proposed. The similar day is selected by the integrated effects of meteorology and time, and the fuzzy membership of the corresponding training sample is obtained by normalized similarity. Using the fuzzy support vector machine to predict the maximum and minimum loads of the forecasting day, then the load is combined with the load curve trend obtained by the linear extrapolation based on the similar day. The simulation results show that the proposed method can improve the predicting accuracy.