选择预报因子的新方法及其对预测的影响分析  

A New Method to Select Predictor and Impact Analysis of Prediction

作　　者：陶凤玲[1] 刘海波 李钊年[1] 王思茹[1] 保广裕[3] 史旺旺 

机构地区：[1]青海大学,青海西宁810016 [2]长江电力三峡电厂,湖北宜昌443133 [3]青海省气象局,青海西宁810001 [4]扬州大学,江苏扬州225009

出　　处：《水电能源科学》2012年第2期187-189,共3页Water Resources and Power

基　　金：江苏高校省级重点实验室开放研究课题基金资助项目(K08016)

摘　　要：针对选择适合预报因子预测负荷时采用相关系数法仅能确定两个随机变量之间线性关系的问题,采用相关概率法选择预报因子构建预测模型,分别考虑了单因素与双因素,并利用最小二乘支持向量机法预测电力负荷。结果表明,该方法精度较高,预测结果较好。According to the problem which correlation coefficient method could only set linear relationship between two random variables when selecting suitable predictor for load prediction,this paper proposes the correlation probability approach for load prediction and establishes the prediction model,which considers single factor and double factors respectively.And then the least squares support vector machine（LS-SVM） is used to forecast load.The results show that the method has high precision and good prediction results.

关 键 词：相关概率法 预报因子 负荷预测 最小二乘支持向量机 

分 类 号：TM714[电气工程—电力系统及自动化]