基于熵权法的公路货运量组合预测

摘要

由于仅采用单一模型进行预测很难得到有效的预测结果，为了充分利用各种预测方法所提供的信息，规避单一模型的局限性，尽可能科学有效地预测公路货运量，在多元线性回归模型、二次指数平滑法以及GM (1,1)模型这三种单一模型的基础上，建立基于熵权法确定权重的组合预测模型。实证结果表明，与单一的预测方法相比，组合预测模型能够降低单一模型的预测风险，减少预测的系统误差，提高预测精度，具有较好的实用价值。%In this paper, we established an entropy weighted combination forecasting model based on the multiple linear regression model, secondary exponential smoothing method and GM(1,1) model. Then through an empirical study, we proved that as compared with the individual models, the combination model had lower forecasting risks and system errors and higher forecasting accuracy.