利用逐步回归预测运输量文档.docVIP

摘 要 本论文主要针对预测需求分析研究方法，利用逐步线性回归的方法研究运量需求问题。为了使经济活动达到顶期的目的，必须对某些经济变量的未来数值作出比较准确的预测，并根据预测的结果采取相应的政策和措施。现实经济生活中也经常需要对客货运量和客货流进行预测。其影响因素都有 (1)国民经济的发展规模和速度。 (2)经济结构的变动(3)基本建设的规模。4)人口因素和生活水平。(5)能源、冶金等工业的规模、速度与布局(6)运输结构的变动等，为了简化模型，我们将影响因素总结称为（1）工业生产总值（2）农业生产总值（3）非商业居民支出，当我们给定一个显著性水平时，通过将变量引入到线性方程中，检测其显著性水平小于给定值时，就说明这个变量对线性方程有显著性影响，反之，则没有显著性影响，将它从线性方程中剔除，接着重新引入变量，再进行检测，原理同上，检测完成后，再对原有变量进行重新检测，看新引入的变量是否对原有变量有影响。重复上述过程，直到引入了所有变量，剔除到不能再剔除为止。这样建立起来的线性回归方程就是对运量需求有显著性影响的回归方程。研究和预测运量的方法有很多，比如：传统方法有 HYPERLINK \l _Toc264276494 1.直接归纳法 HYPERLINK \l _Toc264276495 2.递增率法 HYPERLINK \l _Toc264276496 3.乘车系数法 HYPERLINK \l _Toc264276497 4.产值系数法 HYPERLINK \l _Toc264276498 5.产运系数法；还有多元线性回归的方法。其中为了考虑到显著不显著，我们对比研究了多元线性回归的不足，以及逐步线性回归的优势。同时运量预测在社会生产活动中有着积极重要的作用，为其他社会活动提供了理论依据。 关键词: 运量需求,逐步线性回归, 显著性影响 ABSTRACT This thesis forecast demand analysis method, using the method of linear regression gradually. In order to make economic activities to achieve the purpose of top period of some economic variables, must make the future of comparative accurate numerical prediction, and according to the forecast results adopt corresponding policies and measures. Real economic life also often need to passenger and freight volume and passenger flow forecast. the construction demand factors are: (1) the development of national economy scale and speed. (2) Economic structure change; (3) the basic construction of scale. 4) Population factors and living standards. (5) energy, metallurgy, and other industrial scale, speed and layout (6) transportation structure change etc, in order to simplify model, we will summarize called factors: (1)total industrial production of agricultural production value (2) non-commercial residents spending, when we are given a significant level, through the introduction to the variable linear equation, the significant level is less than a given value, it shows the variables of linear equations have significant influence, and no significant effect, it will be from a linear equation, then remo