电力系统及其自动化学报2019，Vol.31Issue(7)：129-134,6.DOI:10.19635/j.cnki.csu-epsa.000093
电力负荷预测算法比较-随机森林与支持向量机
Comparison Between Power Load Forecasting Algorithms Based on Random Forest and Support Vector Machine
李桂丹 1王佳琦 1靳新悦 1张羽2
作者信息
- 1. 天津大学电气自动化与信息工程学院,天津 300072
- 2. 大唐新能源试验研究院,北京 100052
- 折叠
摘要
Abstract
The comparison conditions of random forest(RF)and support vector machine(SVM)in power load forecast?ing are often ignored,leading to the controversy over prediction accuracy. In this paper,the principles of SVM and RF are studied,and the prediction object and forecasting conditions are analyzed in detail. In addition,three factors includ?ing the algorithm parameters,data set characteristics and weather features are compared. The performances of these two algorithms are statistically compared by changing the forecasting conditions. Results show that SVM algorithm is more sensitive to its own parameters;for the data sets with common changing trends and similar pattern characteristics,the prediction accuracies of both algorithms are obviously better than the other data set;for the weather features used in this paper,temperature and dew point have more impact on load forecasting. However,when both the data set and weather features are the same,statistical analyses show that there is no significant difference in the optimal prediction result between the two algorithms after their parameters are adjusted.关键词
关键字:电力负荷预测/支持向量机/随机森林/参数/数据集/气候特征Key words
power load forecasting/support vector machine(SVM)/random forest(RF)/parameter/data set/weather features分类
信息技术与安全科学引用本文复制引用
李桂丹,王佳琦,靳新悦,张羽..电力负荷预测算法比较-随机森林与支持向量机[J].电力系统及其自动化学报,2019,31(7):129-134,6.
