摘要
以集成经验模式分解(EEMD)与极限学习机(ELM)为理论基础,提出了将其有机结合进行短期电力负荷预测的方法。首先运用EEMD对短期负荷时间序列进行自适应分解,在此基础上针对分解后不同分量的变化特点,建立相对应的ELM模型进行预测,进而将各分量的预测结果重构以得到最终预测结果。通过实例分析表明该模型的预测精度要优于单极限学习机模型、单支持向量机模型和神经网络模型,同时也验证了该模型应用于短期负荷预测的有效性和可行性。
Based on ensemble empirical mode decomposition （ EEMD） and extreme learning machine （ ELM） , a method for short-term load forecasting was proposed. Short-term load time series is adaptively decomposed by using EEMD, based on the characteristics of changes in different components of the decomposition, the establishment of ELM model to the corresponding forecast, and forecasting result of each component reconstruction to get the final prediction results. The practical example analysis shows that the prediction accuracy of the model is superior to single extreme learning machine model, single support vector machine model and neural network model, and it verifies that the validity of the model is applied to short-term load forecast and feasibility.
出处
《贵州大学学报（自然科学版）》
2017年第6期39-42,48,共5页
Journal of Guizhou University:Natural Sciences
基金
福建省中青年教师教育科研项目(JAT170437)
关键词
集成经验模式分解
极限学习机
短期电力负荷预测
支持向量机
神经网络
ensemble empirical mode decomposition
extreme learning machine
short-term power load forecasting
support vector machine
neural network
