基于粗糙集与熵的铁路货物周转量预测

Forecasting railway freight turnover based on rough set and entropy

采用简单熵计算方法对原始数据进行离散,得出决策信息系统。运用粗糙集方法对货物周转量的影响因素进行约简,将约简后条件属性作为解释变量,构建货物周转量多元线性回归模型,运用Eviews软件标定参数,同时对模型的回归显著性进行检验。以2013~2015年货物周转量为例进行预测,得出货物周转量的增长率分别为1.9%,1.0%和-3.8%,周转量分别为29 746.26,30 054.87和29 812.16亿t?km,与实际数据相对吻合度平均为89.05%,表明预测方法具有一定的可靠性。

The simple entropy calculation method was used to discretize the original data,and the decision information system was obtained.The rough set was used to reduce the influencing condition attributes of freight turnover.The rest attributes were taken as explanatory variables,and the explanatory variables and freight turnover were used to establish multiple linear regression model.By using Eviews software to calibrate parameters,the regression significance of the model was tested concurrently.The goods turnover during 2013?2015 was taken as an example to forecast.The predicted growth rate values of goods turnover are 1.9%,1.0%and?3.8%,respectively,whereas the turnover values of goods are 29 746.26,30 054.87 and 29 812.16 tons kilometers,respectively.The average degree of agreement with the actual data is 89.05%,which indicates that the method has certain reliability.