摘要
确定了公路货运量的影响因素分别为GDP、人口数量、社会消费零售总额和农副产品产值,构建了基于模糊线性回归模型的公路货运量预测方法。以延安市公路货运枢纽规划为实例,1995～2004年的货运统计量作为因变量,确定了模型的模糊系数。以2005～2010年的货运统计量作为验证值,分析了模型的拟合精度,并将模糊线性回归模型的预测结果与指数平滑法、灰色模型、弹性系数法3种常见预测方法的预测结果进行比较。研究结果表明:在模糊线性回归模型中,t检验的平均值为0.673 07,说明预测值与实际值差异不显著,模型预测效果较好;4种方法的平均相对误差分别为0.073 1、0.100 3、0.167 8、0.232 9,可见,本文方法误差最小。
The influence factors of highway freight volume were determined, such as GDP, population quantity, the total amount of social consuming retails and the output value of sideline products, and a predictive method of highway freight volume based on fuzzy linear regression model was set up. The highway freight hub planning in Yan＇an City was taken as an example, the statistical freight volumes from 1995 to 2004 were taken as dependent variables, and the fuzzy coefficients of fuzzy linear regression model were determined. The statistical freight volumes from 2005 to 2010 were taken as verified values, and the goodness of fit for fuzzy linear regression model was analyzed. The predictive results among fuzzy linear regression model, exponential smoothing method, grey model and elastic coefficient method were compared. Analysis result shows that in the fuzzy linear regression model, the average value of t test is 0. 673 07, which shows that the difference between predictive value and actual value is not significant, and the prediction effect is better. The average relative errors of four methods are 0. 073 1, 0. 100 3, 0. 167 8, 0. 232 9 respectively, so the error of predictive method is smallest. 5 tabs, 16 refs.
出处
《交通运输工程学报》
EI
CSCD
北大核心
2012年第3期80-85,共6页
Journal of Traffic and Transportation Engineering
基金
国家自然科学基金项目(61101216
51178158)
陕西省交通科技项目(07-10R)
中俄国际道路运输发展研究项目(2011hj-08)
关键词
交通规划
公路货运量
模糊线性回归
预测方法
拟合精度
影响因素
traffic planning
highway freight
fuzzy linear regression
predictive method
grodness of fit
influence factor
