基于灰色关联分析的支持向量机的铁路货运量预测研究
2018-11-02张蕾孙德山张文政王玥
张蕾 孙德山 张文政 王玥

摘 要 采用基于灰色关联分析的支持向量机对铁路货运量进行预测.首先利用灰色关联分析法对影响铁路货运量的因素进行分析处理,然后利用基于高斯核函数的支持向量回归机建立了铁路货运量预测模型.通过分析预测结果可以发现,经过灰色关联分析后的支持向量机模型对复杂的铁路货运量数据有较好地处理能力,且预测相对误差较小.特别地,由于支持向量机的适应性,该模型具有较高的泛化能力,对影响因素较为复杂,样本数量小的预测问题可以提供一定参考.
关键词 铁路货运量预测;灰色关联分析;支持向量机
中图分类号 O213 文献标识码 A
Abstract Using the method of support vector machine which is based on grey correlation analysis to predict railway freight volume. Firstly, using the gray correlation analysis method to analyze the influencing factors of railway freight volume, Secondly using the support vector regression which is based on the Gauss kernel function to establish the prediction model of the volume of railway freight. By analyzing the prediction results, we can find that the support vector machine model which is analyzed by the gray correlation analysis method can process the complex date of the volume of railway freight well, and the relative error of the prediction is relatively smal. Especially, due to the adaptability of support vector machine, the model has a high ability of generalization, and it can provide a reference for the prediction problems with complex factors and small sample size.
Key words applied mathematics; forecast of railway freight volume; grey relational analysis; support vector machine
1 引 言
隨着我国国力日渐强盛,交通运输能力也有了巨大的提升,运输方式逐渐增多,传统运输方式受到猛烈冲击,这种情况在铁路货运市场尤为明显.自2010年起,铁路货运量开始逐渐下滑.2011年,我国铁路货运量为393263亿吨,占全国总货运量的10.63%,而到了2015年,铁路货运量为335801亿吨,同比降低10.53%,几乎跌至6年前的水平.面对越来越严峻的货物运输市场,铁路货运在面临着巨大的挑战的同时也充满了新的机遇.铁路货运管理部门如果想要抓住机遇,焕发生机,就需要更加准确掌握铁路货运未来的发展趋势.
影响铁路货运量的因素万缕千丝,这些因素对货运量的作用机制又很难用精确的数学语言来表示,这就使数学预测模型难以建立.传统的预测方法有:线性回归法、时间序列法、状态空间法和指数平滑法等;……
