摘要
采用灰色关联分析法筛选出江西省铁路货物周转量的主要影响因素,在此基础上建立了BP神经网络预测模型,并采用多元线性回归模型、二次指数平滑法、灰色GM(1,1)模型分别对江西省铁路货物周转量进行预测,再对结果进行比较和误差分析。研究表明,BP神经网络模型预测精度明显高于其它三个模型,平均误差为0.76%,可用于实际预测。
On the basis of selecting the main influence factors by gray correlation analysis, BP neural network model was established,and the freight turnover in Jiangxi was forecasted and comparatively analyzed by using multiple linear regression, two exponential smoothing and grey GM(1,1) model. The result shows that BP neural network is obviously more accurate than other prediction methods and can be applicable to practice, its mean error is 0.76%.
出处
《齐齐哈尔大学学报（自然科学版）》
2014年第1期50-53,共4页
Journal of Qiqihar University(Natural Science Edition)
关键词
货物周转量
灰色关联分析
BP神经网络
预测
freight turnover
gray correlation analysis
BP neural network
prediction
