基于灰色支持向量机的季节型负荷预测方法  

Seasonal load forecast methods based on grey Support Vector Machine

作　　者：牛东晓[1] 谷志红[1] 王会青[2] 王维军[1] 

机构地区：[1]华北电力大学工商管理学院,北京102206 [2]山西大学经济与工商管理学院,山西太原030006

出　　处：《华东电力》2007年第6期1-5,共5页East China Electric Power

基　　金：国家自然科学基金资助项目(70671039)

摘　　要：季节型负荷具有增长性和波动性的二重趋势,并且呈现出复杂的非线性特征,同时又受到多种随机干扰因素的影响,难以用单一的预测模型做出准确的预测。提出一种基于粗糙集的灰色支持向量机预测系统,将该系统应用于季节型负荷预测中,与单一的GM(1,1)方法和BP神经网络法相比,得到了较高的预测精度。Because of the dual trends （increase and fluctuation） and their complex nonlinearity, the seasonal load which is also subject to multiple stochastic interference factors is difficult to be forecasted with single model. To solve the problem, a forecast system using rough set-based grey Support Vector Machine is proposed and was applied to seasonal load forecast. The system is accurate in forecast in comparison with the single GM （ 1,1 ） method and BP neural network method.

关 键 词：电力系统 粗糙集理论 剥离因子 灰色预测模型 支持向量机 季节型负荷预测 

分 类 号：TM715[电气工程—电力系统及自动化]