| Cluster analysis has great value in social science research. In the absence of prior knowledge, Cluster analysis can decompose the research object (sample or index) into a plurality of classes according to their degree of intimacy in the nature, so that objects have a high similarity in the same class, objects with large differences in different classes. Cluster analysis can identify data distribution structure and the degree of similarity between the data sets.In the social problems decision-making process, we often encounter some problems:different regions, industries, different levels of income groups, these can not be generalized when we make decisions, we require classification de-cision. To research the classified data, we often use cluster analysis methods to identify data distribution structure and similarity measurement so that we can solve the problem properly and efficiently.Clustering algorithm is used so widely. In business, it can help market researchers classify different consumer groups in a large number of consumer data, and describe the characteristics of various consumer groups in different consumption patterns; In social economic development, it can cluster analysis the regional economic and social development level, and classify and evaluate the regional and national economy level; In biology, it can classify a variety of plants or animals, classification of genes and proteins in order to obtain the understanding of the composition of the population structure. Clustering algorithms can independently obtain data distribution situation, observe the characteristics of each cluster classification, and re-analysis of these character- istics node.After years of development, the regional higher education has developed very rapid in our country, and they transport a large number of highly educated graduates for the state and community. However, because of different parts of our country showing uneven economic development, structure and layout of the various regions of university innate lead that the development of each region in our country is not in the same horizontal line, so our regions to 31 higher levels have a certain degree of difference. This paper makes use of cluster analysis method for our country regional higher education development-related data sorted and explores the city higher education forward the development process in all regions of difference, law, structure and characteristics. In this paper,the cluster analysis based on the development of general higher education in vari-ous regions, can clear the differences and features of regional higher education development situation. And it is conducive for the management and decision-making departments to grasp the overall status of the development of higher education in China from the macroscopic, classification to formulate relevant policies, better guidance and planning the overall healthy development of the higher education in China.This paper introduces the main theory of clustering algorithm, introducing the basic idea of the various clustering algorithms and the applicable data set types, according to the sample or index to explore how to choose a reasonable clustering algorithm. To regions of higher education data sets in China case, we make use of MAT LAB and SPSS softwares to achieve clustering analysis of data set. According to the results of the algorithm, clear differences between the characteristics of each area of general higher education. According to the characteristics of politics, economic, culture, geography and humanity of the regions, recommendations targeted the development of relevant policies, development strategy of expanding the difference when the difference in various regions of under development, the implementation of the development strategy of balanced when difference is too large, in order to facilitate adjustment of the overall healthy development of the higher education, at the same time reflect the practical value of clustering analysis in social, political and economic fields. |
PDF Full Text Request