摘要
针对目前超短期风速预测精度不高的问题,提出了一种改进样本加权的SVM超短期风速预测方法。对样本加权中基于距离函数的时间序列相似性度量方法进行改进,在欧式距离的基础上,加入区间变化趋势相似度函数,将欧氏距离和趋势相似度函数按权值组合,构造了新的相似性度量函数。对训练样本进行相空间重构,基于样本相似性因素对训练样本进行加权,建立加权SVM超短期风速预测模型。分别建立随机森林、梯度提升树、SVM以及改进加权SVM超短期风速预测模型,研究表明,对SVM进行改进样本加权后,可以将预测误差从7.61%降为7.46%,有效降低了超短期风速预测误差,验证了该方法的有效性。
Aiming at the problem that the prediction accuracy of ultra short term wind speed is not high at present, an improved sample weight SVM method for ultra short term wind speed prediction is proposed. Based on the distance of Euclidean distance, the similarity function of interval trend is added, and the Euchdean distance and trend similarity function are combined by weight to construct a new similarity measure function. The training sample is reconstructed by phase space, and the training sample is weighted Based on the sample similarity factor to estabhsh the model of weighted SVM ultra short wind speed prediction. The re- suits show that the prediction error can be reduced from 7.61% to 7.46% after the improved weighting of SVM, which can effec- tively reduce the ultra short term wind speed prediction. The results show that the proposed method can reduce the prediction er- ror of SVM Error, which verifies the effectiveness of the method.
作者
张瑞成
田新
ZHANG Rui-cheng, TIAN Xin （College of Electrical Engineering,North China University of Science and Technology, Tangshan 063210, China）
出处
《电脑知识与技术》
2017年第9期211-214,共4页
Computer Knowledge and Technology
基金
河北省自然科学基金资助项目（F2014209192）
华北理工大学杰出青年基金资助项目（JP201301）
关键词
样本加权
相似性
相空间重构
支持向量机
风速预测
sample weighting
similarity
phase space reconstruction
SVM
windspeed prediction
