摘要
首先分析得出与功率有关的变量,然后根据互信息理论对变量通过最大相关最小冗余的原则进行特征选取,挖掘特征与功率之间的联系,最后根据这种联系对功率数据进行补齐。利用该方法对吉林省某风电场进行算例验证,结果表明,随机缺失的数据补齐之后的准确率高于连续缺失后补齐的结果,而且基于相关向量机模型补齐的输出功率的结果误差减小,准确率提高。
The power-related variables was analyzed, then the variables were selected according to mutual information theory and the principle of maximum related minimum redundancy （MRMR）, the relation between features and power was excavated, finally the power data were polished according to the relation. The method was used to verify a certain wind farm in Jilin province as numerical example. The results indicate that the accuracy of polished random missing data are higher than that of polished continuously missing data, and the error of polished output power results based on relevance vector machine model is reduced, the accuracy is improved.
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2017年第4期938-944,共7页
Acta Energiae Solaris Sinica
基金
国家重点基础研究发展(973)计划(2013CB228201)
国家自然科学基金(51307017)
吉林省科技发展计划(20140520129JH)
吉林省产业技术与专项开发项目(2014Y124)
关键词
风电功率
数据补齐
最大相关最小冗余
互信息
相关向量机
wind power
data polishing
maximum related minimum redundancy
mutual information
relevance vectormachine
