基于数据挖掘的负荷预测

Load forecast based on data mining

简要叙述了数据挖掘技术及其在电力系统中的应用情况。提出了一种基于数据挖掘中时间序列相似性研究的电力负荷预测方法。通过采用序列分段平均值技术降维,结合滑动窗口和MBR方法实现子序列相似查询,并利用R+-树作为多维索引结构提高检索效率。实验结果证明提出的方法是行之有效的。

Data mining technique and its applications in power system are briefly introduced and a electrical load forecast method is presented,which is based on the similarity research of time-se-ries in data mining.Piecewise average approximation is used to reduce dimensionality and slipping window combined with MBR to implement subsequence matching.R + -tree is taken as multi-dimen-sional structure to improve indexing efficiency.The experimental results prove this method is effec-tive.