摘要
由于电网工程塔材价格具有非线性和非平稳性特征,导致其价格预测难度大、预测精度低,针对这一问题,建立了EMD-ARMA预测模型。利用经验模态分解(EMD)将历史价格分解为平稳的、周期波动的若干价格分量,并以此作为输入,利用自回归滑动平均模型(ARMA)对各分量进行价格预测,最后将各预测分量叠加得到预测值。以Q345塔材的历史数据为样本,通过与ARMA的预测结果进行对比及误差分析,验证了本文采用的EMDARMA预测方法能够有效提高电网工程塔材价格的预测精度,对于工程造价管控和设备材料招投标具有一定的参考价值。
The tower material price of grid project is nonlinear and non-stationary,so the price is difficult to forecast,and prediction accuracy is low.According to this,the EMD-ARMA model is built.The historical price is decomposed into some smooth and periodic fluctuation components by empirical mode decomposition （EMD）.Taking the components as input parameters,the components＇price is forecasted by auto-regressive and moving average model （ARMA）.Finally,the forecasted price is obtained through the superimposition of each forecasted component.Based on the history data of Q345 tower material,the example analysis shows the accuracy of forecasted result by EMD-ARMA is higher than that of the result by ARMA.The forecast accuracy of the tower material price is improved effectively by EMD-ARMA.
出处
《陕西电力》
2013年第11期70-74,共5页
Shanxi Electric Power
关键词
电网工程塔材价格
经验模态分解
自回归滑动平均模型
预测
tower material price for grid project
empirical mode decomposition （EMD）
auto-regressive and moving average model （ARMA）
forecasting
