基于快速K均值聚类的经济水平与货运量模型

Economy level and freight model utilizing quick K-means cluster method

应用快速K均值聚类方法对我国各地区工业生产水平、单位生产的货运量进行聚类分析。根据各省市在分类成员表中的类别归属，对横截面分类数据作了进一步的回归分析。选取工业生产水平和货运量的对数为2个聚类因子，得出了5个聚类。通过定性和定量相结合的方法得出我国各地区的经济水平与货运量增长速度之间的定量关系模型，模型的相关系数介于0．901－0．982，置信度较高。

The cluster analysis for the industry production level and the freight volume of different regions in China were performed using the quick K-means cluster method to gain the classification in the cluster membership. A further regression analysis for the sectional classification data was made based on the classification data in the cluster membership of ,different provinces and cities in China. Five clusters were achieved using the logarithms of the industry product level and the freight volume as 2 cluster factors. The model for the quantitative relations between the growth rates of the regional economical level and the regional freight volume was established through the combined qualitative and quantitative analyses. The correlation coefficient of this model ranges from 0. 901 to 0. 982,indicating its superior internal consistency and high confidence level.