| 引用本文: | 李军,常燕芝.基于KPCA-KMPMR的短期风电功率概率预测[J].电力自动化设备,2017,37(2): |
| LI Jun,CHANG Yanzhi.Short-term probabilistic forecasting based on KPCA-KMPMR for wind power[J].Electric Power Automation Equipment,2017,37(2): |
|
| 摘要: |
| 针对短期风电功率概率预测，提出一种基于核主成分分析(KPCA)与核最小最大概率回归机(KMPMR)相结合的方法。 KPCA方法可对数据进行预处理，在特征空间中有效提取模型输入的非线性主元；KMPMR方法在仅需假定产生预测模型的数据分布的均值与协方差矩阵已知时，将最小最大概率分类机(KMPMC)的分类超平面看作预测模型的输出，可最大化模型的输出位于其真实值边界内的最小概率。实验结果表明，所提方法在预测精度上优于现有的预测方法，并能提供预测误差的分布范围。 |
| 关键词: 核主成分分析 核最小最大概率回归机 风电功率 概率预测 |
| DOI：10.16081/j.issn.1006-6047.2017.02.004 |
| 分类号:TM614 |
| 基金项目:国家自然科学基金资助项目(51467008) |
|
| Short-term probabilistic forecasting based on KPCA-KMPMR for wind power |
|
LI Jun, CHANG Yanzhi
|
|
School of Automation & Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China
|
| Abstract: |
| A probabilistic forecasting method of short-term wind power based on the combination of KPCA (Kernel Principal Component Analysis) and KMPMR(Kernel Minimax Probability Machine Regression) is proposed, which applies KPCA to pre-process the data for the effective extraction of the nonlinear principal component from the feature space as the input of forecasting model. Assuming the mean and covariance matrix of the distribution which generates the forecasting model are known, the KMPMC method regards the classification hyperplane of KMPMC as the output of forecasting model for maximizing the minimum probability of the model output within the boundary of its true value. Experimental results show that, the proposed method has better forecasting accuracy than the existing forecasting methods and it can provide the probability distribution of forecasting error. |
| Key words: kernel principal component analysis kernel minimax probability machine regression wind power probabilistic forecasting |
