基于集合经验模态分解和特征选择极端学习机的风速预测

Short-term Wind Speed Forecasting Using Ensemble Empirical Mode Decomposition and Extreme Learning Machine with Feature Selection

提出一种集合经验模态分解、随机森林和极端学习机建模的短期风速预测方法。首先,引入集合经验模态分解将原始风速数据分解成代表不同波动特征的分量,剔除不规则的分量;然后,对保留分量逐一建模,构建随机森林特征选择算法,根据重要性来提取模型的特征输入;最后,建立基于特征选择和极端学习机的风速分量预测模型进行预测,综合分量预测结果得出最终预测结果。

This paper proposes a hybrid model for short-term wind speed forecasting using ensemble empirical mode decomposition(EEMD),random forest(RF)and extreme learning machine(ELM).Firstly,EEMD is introduced to decompose the original wind speed signal into several sub-series,representing different fluctuation information.Sub-series which represents irregular fluctuation is eliminated before modeling.Then each sub-series is modeled by using ELM with feature selection respectively.RF is adopted to select the inputs according to the importance of each feature.Finally,the forecasting results of the each sub-series are summed to obtain the predicted values of wind speeds.A case study is carried out to verify the validity of the proposed model.