基于Shapley值组合模型的我国快递业务量预测

Prediction of Express Delivery Volume in China Based on Shapley Value Comprehensive Model

快递业务量的预测是我国邮政业制定发展策略的重要前提。通过采用灰色GM(1,1)预测、多项式趋势外推预测、指数平滑预测及Shapley值分配权重的组合预测模型对我国2012-2021年快递业务量进行分析,对比预测结果与平均相对误差,证实了基于Shapley值法的组合预测模型在我国快递业务量中的预测精度高于三种单一预测模型,具有科学性。利用此模型对我国“十四五”期间快递业务量进行预测,为政府、邮政管理部门及快递相关企业制定发展策略提供参考。

Through the comprehensive forecast model,which includes grey GM(1,1)forecast,polynomial trend extrapolation forecast,exponential smooth forecast,and combined with Shapley value for weight distribution,we analyzed the express delivery volume in China from2012 to 2021.By comparing the prediction results and the average relative errors,we confirmed that the model combining the above is more scientific,and its forecasting accuracy for express delivery volume more accurate than any of the three individual models.Next,we used the combinational model to forecast the express delivery volume of China during the"14th Five-Year Plan",providing reference for the government,postal administration departments and related enterprises to formulate more accurate development strategies.