基于神经网络算法的公路货运量预测方法

以陕西省为例,运用灰色关联分析法确定公路货运量的影响因素分别为地区生产总值、第一产业增加值、第二产业增加值、工业增加值、人均地区生产总值、全社会固定资产投资和社会消费品零售总额.将所确定的因素作为公路货运量的预测指标,建立基于BP神经网络的公路货运量预测模型,并对模型进行应用测试.结果表明:该模型具有较高的精度,最大误差为5.3%,可以提高公路货运量预测的准确度,为我国公路货运量的预测研究提供方法支撑.

Forecast Method of Road Freight Traffic Based on BP Neural Network

Shanxi Province was taken as an example for the road freight traffic forecasts by using gray correlation method. The predictors are GDP, the first industry, the secondary industry, industrial added value, per capita GDP, total fixed asset investment and the total retail sales of social consumer goods. The prediction model of road freight traffic is established on base of BP neural network, and then verified with tests. The results show that road freight traffic can be predicted accurately by the model based on BP neural network, and the maximum error is less than 5.3%. It can improve the forecast ability of road freight traffic and provide a method for road freight traffic.