城市日用水量预测的部分线性自回归模型.pdfVIP

2 8 2 V o l28N o2 2009 6 Journa l o f W uhan P olytechn ic U n iversity Jun. 2009 : 2009 DO I: 10. 3969 / j. issn. 2009. 02. 030 城市日用水量预测的部分线性自回归模型 , , ( , 4 30205 : 根据城市用水量影响因素及特点, 针对线性回归模型与线性自回归模型误差较大的 缺点, 建 了城市日用水量的部分线性自回归预测模型, 其中线性部分考虑日用水量, 非线性 部分考虑当天的最高温度, 该模型综合了非参数回归模型与线性自回归模型的优点, 因此在拟 合与预测精度上比线性回归与线性自回归模型有所提高, 证明该方法在城市日用水量预测中 是有效的 : 部分线性自回归模型; 日用水量; 变量选择 : O 29; TU 99 1. 31 : A Pa rtia lly linea r AR mode l of urban da ily w ater demand CH EN Zhanbo, GENG Zh ilin, YAN G K eling ( Departm ent of S tatistics and m athem at ics, H ubei Un iversity of E conom ics, W uhan 430205 Ab stract: A ccording to the affecting factor and the character of urban da ily w ater dem and, the partially linear A R m ode l of urban daily w ater dem and is set up. T he linear aspect takes into account the w ater dem and and the non linear aspect takes into account tiptop temperature. T his mode l not only has the m erit of regressive m odel but also has the auto regressive m ode l m erit. Com pared w ith the form er three m odels, the smi ulation and forecasting resu lt o f this m ethod is better, so the m ethod is effect ive. Key wo rds: partia lly linear AR m ode;l w ater da ily dem and; variab le selection , , , , , , , [ 1] ( [ 2] [ 3] ( 1998 , ( 2001 , ( 2004 , , [ 4] ( 2004 , , , , ,