摘要
为有效提高造价预测的准确性,建立EEMD-SVM模型,将造价的历史数据平稳化处理后作为输入,利用支持向量机(SVM)对各造价分量进行预测,经叠加得到输电线路工程的造价预测值。以某地区110k V输电线路工程造价的历史数据为样本,经EEMD-SVM预测,结果经过平均相对误差(MAPE)、均等系数(EC)以及均方根误差(RMSE)检验,误差较小。因此,该模型在一定程度上提高了输电线路工程造价的预测精度,对于提升造价管理的精益化水平具有一定的参考价值。
In order to improve the accuracy of cost forecasting,the paper builds the EEMD-SVM model,which takes the history cost datas as input after stationary processing,and forecasts each cost components by Support Vector Machine（SVM）,gets the cost forecasting values of transmission line engineering by the superimposing. Taking the history cost datas of the 110 k V power transmission line engineering as sample,by EEMD-SUM forecasting,tests the forecasted results through MAPE、EC and RMSE methods,and the error is less. Therefore,the model improves the accuracy of cost forecasting of the transmission line engineering,and provides references for improving the level of lean cost management.
作者
卢艳超
温卫宁
王晓晖
徐丹
LU Yanchao WEN Weining WANG Xiaohui XU Dan(State Power Economic Research Institute, Beijing 100029, China)
出处
《工程经济》
2017年第6期5-9,共5页
ENGINEERING ECONOMY
关键词
输电线路工程
造价预测
支持向量机
预测精度
transmission line engineering
cost forecasting
Support Vector Machine（SVM）
forecasting accuracy
