基于并行共享挖掘算法的电力负荷预测  

Power Load Forecasting Based on Parallel Shared Mining Algorithm

作　　者：赵文硕[1] 谢萍[1] 王颖[1] 李彦[1] 廖一鸣[1] 

机构地区：[1]华北电力大学控制与计算机工程学院,北京102206

出　　处：《计算机与数字工程》2015年第2期178-182,共5页Computer & Digital Engineering

摘　　要：影响电力负荷的因素有很多,用于电力负荷预测的数据随着电网规模的扩大也越来越多。传统预测方法不能通过挖掘大量数据,有效地建立电力负荷预测模型。论文结合并行共享数据挖掘技术,综合分析了基于HADOOP的并行共享决策树算法(PSDT)以及SLIQ算法,提出了一种新的方法来构建电力负荷预测模型。试验结果表明,使用该方法构建预测模型实用性较强,并且有良好的扩展性。There are many factors that affect the power load ,and data for electric power load forecasting is also more and more .Traditional forecast methods can’t effectively establish power load forecasting model through mining large data . This paper puts forward a new method to establish a power load forecasting model ,combining the parallel shared data mining technology and analysising the parallel shared decision tree based on HADOOP algorithm（PSDT ） and the SLIQ algorithm . The experimental results show that ,using this method to establish power load forecasting model has better practicability , and has good scalability .

关 键 词：数据挖掘 负荷预测 HADOOP 并行共享决策树 

分 类 号：TM714[电气工程—电力系统及自动化]