摘要
提出燃气日负荷预测模型选取和变量选取的原则,给出数据处理的方法。以西南某市为例,依据该地天然气公司实际数据,给出基于支持向量回归(SVR)的燃气日负荷预测模型的创建流程,进行日负荷预测,将日负荷预测值与实际值进行比较,分析相对误差产生的原因。探讨运用SVR模型建设信息化系统的相关问题。
The principles of model selection and variable selection for daily gas load forecast are proposed,and the method of data processing is given.Taking a city in Southwest China as an example,according to the actual data of the local natural gas company,the creation process of daily gas load forecast model based on support vector regression(SVR)is given,the daily load forecast is carried out,the daily load forecast value is compared with the actual value,and the causes of relative errors are analyzed.The related problems of using SVR model to build information system are discussed.
作者
陈效吉
朱喜楠
张佳妮
CHEN Xiaoji;ZHU Xinan;ZHANG Jiani
出处
《煤气与热力》
2021年第5期I0004-I0007,I0044,共5页
Gas & Heat
关键词
燃气日负荷预测
支持向量回归
变量选取
相对误差
核函数
daily gas load forecast
support vector regression(SVR)
variable selection
relative error
kernel function
