摘要
为了提高短期风速预测精度,提出一种变权系数的支持向量机组合风速预测模型。选择基于不同核函数的支持向量机作为单项预测模型以保证单项模型之间的差异性,对核参数用粒子群算法寻优选取以保证各单项模型的精确性。组合预测方法采用以预测误差平方和最小为准则的可变加权系数组合预测方法,以计算各单项模型在风速预测不同时刻的权系数。仿真实验表明,所建立的变权组合预测模型在短期风速预测上具有良好的预测效果,预测精度优于各单项模型和固定权系数的组合模型。
In order to improve the prediction accuracy of short-term wind speed, a combination forecasting model based on support vector machine (SVM) with variable-weight coefficient was presented. The support vector machine with different kernel parameters was chosen as the single forecasting model to guarantee the difference between single models. The selection of kernel parameters was optimized using particle swarm algorithm to ensure the accuracy of each individual model. The combination forecasting method satisfying the variable weight coefficient with the least square sum of forecasting errors was used to calculate the weight coefficient of the each individual model at different time of wind speed forecast. The simulation experimental results show that the proposed combination forecasting model with variable weight coefficient has better forecast result for short term wind speed prediction, and the forecast accuracy is higher than the each single model and combination model with fixed-weight coefficient.
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2017年第6期1510-1516,共7页
Acta Energiae Solaris Sinica
基金
中央高校基本科研业务费专项资金(2014MS139)
关键词
风速预测
支持向量机
组合预测
变权系数
混沌相空间重构
wind speed prediction
support vector machine
combination forecasting
variable weight coefficient
chaosphase space reconstruction