组合模型在水库泥沙淤积量预测中的应用  

Application of combining forecast model in the prediction of reservoir sedimentation amount

作　　者：高洪波[1] 洪为善 

机构地区：[1]江苏城市职业学院科研处,江苏南通226006 [2]长江水利委员会汉江水文局,湖北丹江口442700

出　　处：《南水北调与水利科技》2014年第6期200-203,共4页South-to-North Water Transfers and Water Science & Technology

摘　　要：基于指数平滑、模糊移动、线性回归、灰色预测四种单一模型,建立了组合预测模型,并根据某水库1997年-2012年泥沙淤积量实测数据,构建了该水库泥沙淤积量组合预测模型。实证研究结果表明,组合预测模型在有效综合各单一模型的有用信息方面具有较强的优势,模拟结果平均绝对误差百分比为2.539 3%,均高于四种单一模型的预测精度,因此组合模型具有较好的实用价值。A combining forecast model was developed based on four individual models,including the exponential smoothing,fuzzy moving,linear regression,and grey forecast models.Moreover,a combining forecast model was developed for a reservoir based on the measured sedimentation amounts from 1997 to 2012.The results showed that the combining forecast model has advantages in combining the useful information from individual models,and the mean absolute error percentage is 2.539 3％ for the combining forecast model,which has higher prediction accuracy than the individual models.Therefore,the combining forecast model has practical values.

关 键 词：指数平滑 线性回归 模糊移动平均法 灰色模型 组合模型 水库泥沙淤积量 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]