Research On The Logistics Development Levels Of 17 Cities In Shandong Province Based On Cluster Analysis

Logistics is an important link between production and consumption, city and surrounding areas. An advanced level of logistics development can greatly promote the city’s industrial upgrading and transformation, improve economic efficiency, and provide continuously power for development of a city or even a country. It has great influences on the socio-economic development. Against the backdrop many provinces and regions have introduced various related incentives, supporting the construction of logistics infrastructure is constantly being improved. In this process, if we can’t make a reasonable evaluation for the scale of demand and infrastructure capacity, but blindly development, that will cause some impediments on logistics industry.Compared with areas which have developed logistics industry, such as Shanghai, Jiangsu, Zhejiang and so on, the logistics industry of Shandong Province is still at a low level. The logistics development between various cities in Shandong province is quite different, distribution of resources is uneven, and the development of supply and demand is unbalanced. In this paper, we refer to the research achievements of domestic and foreign scholars on the theory and practice of the regional logistics development. Based on the actual status of logistics development of the 17 cities in Shandong province, we collected information and data. Then we construct regional logistic development evaluation system which is suitable for Shandong province through the perspective of supply and demand. This index system can be more comprehensive and rational reflect the key factors in the development of cities in Shandong province.Firstly, we build logistics evaluation index system according to the current logistics development situation in Shandong province on the basis of a large amount of investigation and relevant research literatures. We dig the logistics development balance within the system from aspects of supply and demand, and this is a relatively new method.Secondly, for cross section data and panel data, respectively, we use different methods of clustering analysis. For Logistics development section data in 2013 in Shandong province, we use the principal component analysis (PCA) for data dimension reduction, then use a new type of graph clustering method (QCM) to do clustering operation. All the 17 cities in Shandong province were divided into three levels, comparing three clusters we made analysis on the results. Besides, some reasonable and scientific recommendations on its development are raised.Finally, biclustering methods were innovatively used to analyze regional logistics which are widely used in bioinformatics. For the logistics dynamic data in the past twelve years in Shandong province, we use time factor discounting methods to reduce the data dimension in the time dimension, then do the cluster analysis from city and index prospects at the same time. The results show that all the 17 cities were divided into four levels, analyzed the similarity between the cities within each level from the perspective of not only the city but also from the perspective of index. At last we put forward reasonable suggestions based on the results. The g-CLUTO program provides visualization matrix and visual hill diagram makes the clustering results more intuitive.