Research Of Urban Competitiveness Of Gansu Province Based On Multivariate Analysis

Nowadays,with the rapid development of economy,the competition between cities becomes more and more fierce.And in the face of fierce competition among the cities,how to evaluate city competitiveness,improve the city's comprehensive competitiveness are the main content of the urban competitiveness research.At home,the study of urban competitiveness in this area is relatively late,the evaluation model and index selection standards do not appear to have reached a consensus,and city competitiveness evaluation index involving many variables.How to analyze the multivariate data to comprehensive evaluation of urban competitiveness,is the research contents of this paper.First in this paper,the 12 cities in gansu province as sample,in the state of around 2008-2012 index five groups of data based on the empirical analysis,using principal component analysis to simplify data,dimension,visualization,and its results from lateral to do comprehensive evaluation and ranking from city to city,from longitudinal summary analysis of city development trend.Secondly,based on principal component analysis,fuzzy c-means clustering algorithm for five groups of data clustering analysis respectively.In the urban comprehensive competitiveness of 12 cities in Gansu province is divided into three categories: Lanzhou to the first class,Jiayuguan,Jinchang,the Jiuquan to the second category,Tianshui,Wuwei,Baiyin,Zhangye,of out,Qingyang,Dingxi,Gansu for the third class,with 12 years of data as the base analysis of dominant factors,and the classification results visualization.After the final dimension with principal component analysis data,the application of spectral clustering and particle swarm optimization(pso)algorithm respectively for the classification of the data visualization,and compared with the fuzzy c-means clustering analysis results,the mutual authentication,three methods of data classification results are basically identical,but from the point of theory and empirical analysis of visual interface,spectral clustering and according to the analysis of the particle swarm optimization(pso)algorithm for urban competitiveness.