基于K-prototype与SVM的短期电力负荷预测
|
摘要
电力负荷受多种不确定因素的影响,所以通过使用一种聚类方式选取相似日来进行短期电力负荷预测,以提高预测精度和速度。K-prototype聚类算法继承了k均值算法所以具有k均值优点。通过建立映射数据库将气温、湿度、天气、星期类型、季节等不同量纲的值通过无量纲化处理,使得不同的量纲在数值上具有可比性。然后运用聚类方法选取预测日的相似日,借助SVM模型对相似日样本进行预测,实现更加精准预测的目的。利用某地区真实数据进行验证,结果表明该方法可以有效地提高短期电力负荷的预测精度。
The electrical load is affected by many uncertain factors, so short-term electric load forecasting is carried out by using a clustering method to select similar days to improve the prediction accuracy and speed. The K-prototype clustering algorithm,which comes from the k-means algorithm,has the advantage of k-means. By establishing a mapping database including data of temperature,humidity,weather,day type,and season,the values of different dimensions are processed by dimension lessness to be comparable. After selecting the similar day of the forecast day by the clustering algorithm,it applies SVM model to predict the similar day samples to get more accurate prediction. The verification of real data in a certain area shows that the method can effectively improve the prediction accuracy of short-term electrical load.
The electrical load is affected by many uncertain factors, so short-term electric load forecasting is carried out by using a clustering method to select similar days to improve the prediction accuracy and speed. The K-prototype clustering algorithm,which comes from the k-means algorithm,has the advantage of k-means. By establishing a mapping database including data of temperature,humidity,weather,day type,and season,the values of different dimensions are processed by dimension lessness to be comparable. After selecting the similar day of the forecast day by the clustering algorithm,it applies SVM model to predict the similar day samples to get more accurate prediction. The verification of real data in a certain area shows that the method can effectively improve the prediction accuracy of short-term electrical load.
引文
[1]康重庆,夏清,刘梅.电力系统负荷预测[M].北京:中国电力出版社,2007.
[2]刘琪琛,雷景生,郝珈玮,等.基于Spark平台和并行随机森林回归算法的短期电力负荷预测[J].电力建设,2017,38(10):84-92.
[3]吴润泽,包正睿,王文韬,等.Hadoop架构下基于模式匹配的短期电力负荷预测方法[J].电工技术学报,2018,33(7):1542-1551.
[4]盖姝.基于改进支持向量机的短期电力负荷预测研究[D].保定:华北电力大学,2013.
[5]魏立兵,赵峰,王思华.基于人群搜索算法优化参数的支持向量机短期电力负荷预测[J].电测与仪表,2016,53(8):45-49,74.
[6]Pitt B D,Kitschen D S.Applications of data mining techniques to electric load profiling[C].Manchester:University of Man chester Institute of Science and Technology,1999:131-136.
[7]孙海蓉,谢碧霞,田瑶,等.基于数据的SecRPSO-SVM短期电力负荷预测[J].系统仿真学报,2017,29(8):1829-1836.
[8]聂品磊,费东,王宏杰,等.基于EMD-BP神经网络的短期电力负荷预测[J].化工自动化及仪表,2016,43(3):305-307,332.
[9]陈亚,李萍.基于神经网络的短期电力负荷预测仿真研究[J].电气技术,2017(1):26-29.
[10]李如琦,杨立成,苏媛媛,等.基于相似日和蚁群优化LS-SVM的短期电力负荷预测[J].现代电力,2008(2):33-37.
[11]Vapnik V N.An Overview of Statistical Leaning Theory[J].IEEE Trans.on Neural Netw,2008,10(5):988-999.
[12]Huang Z X.Clustering large data sets with mixed numeric and categoricalvlues[C].Proceedings of the 1st Pacific-Asia Knowledge Discovery and Data Mining Conference.Singapore,1997:21-34.
[13]Cristianini N,Taylor J S.支持向量机导论(李国正,王猛,曾华军)[M].北京:电子工业出版社,2004:98-107.
[14]康重庆,程旭,夏清,等.一种规范化的处理相关因素的短期负荷预测新策略[J].电力系统自动化,1999(18):32-35.
[2]刘琪琛,雷景生,郝珈玮,等.基于Spark平台和并行随机森林回归算法的短期电力负荷预测[J].电力建设,2017,38(10):84-92.
[3]吴润泽,包正睿,王文韬,等.Hadoop架构下基于模式匹配的短期电力负荷预测方法[J].电工技术学报,2018,33(7):1542-1551.
[4]盖姝.基于改进支持向量机的短期电力负荷预测研究[D].保定:华北电力大学,2013.
[5]魏立兵,赵峰,王思华.基于人群搜索算法优化参数的支持向量机短期电力负荷预测[J].电测与仪表,2016,53(8):45-49,74.
[6]Pitt B D,Kitschen D S.Applications of data mining techniques to electric load profiling[C].Manchester:University of Man chester Institute of Science and Technology,1999:131-136.
[7]孙海蓉,谢碧霞,田瑶,等.基于数据的SecRPSO-SVM短期电力负荷预测[J].系统仿真学报,2017,29(8):1829-1836.
[8]聂品磊,费东,王宏杰,等.基于EMD-BP神经网络的短期电力负荷预测[J].化工自动化及仪表,2016,43(3):305-307,332.
[9]陈亚,李萍.基于神经网络的短期电力负荷预测仿真研究[J].电气技术,2017(1):26-29.
[10]李如琦,杨立成,苏媛媛,等.基于相似日和蚁群优化LS-SVM的短期电力负荷预测[J].现代电力,2008(2):33-37.
[11]Vapnik V N.An Overview of Statistical Leaning Theory[J].IEEE Trans.on Neural Netw,2008,10(5):988-999.
[12]Huang Z X.Clustering large data sets with mixed numeric and categoricalvlues[C].Proceedings of the 1st Pacific-Asia Knowledge Discovery and Data Mining Conference.Singapore,1997:21-34.
[13]Cristianini N,Taylor J S.支持向量机导论(李国正,王猛,曾华军)[M].北京:电子工业出版社,2004:98-107.
[14]康重庆,程旭,夏清,等.一种规范化的处理相关因素的短期负荷预测新策略[J].电力系统自动化,1999(18):32-35.