A Clustering Analysis Of Chinese Urbanization Level Based On A Nonparametric Methodology

Urbanization has been a study focus for a long time. Because theregional income gap enlarges along with the economic development, thusthe urbanization level experiences a corresponding change.The traditionaldefinition of urbanization is the proportion of urban population to totalpopulation in a region, it is easy to carry. While the number of urbanpopulation is enslaved to the household register management policy ofthe region and the traditional index for urbanization lacks of horizontalspatial comparison and lengthways comparison, so the traditionaldefinition may probably not show the exact urbanization level of a region.A key task in front of us is to set up a series of efficient and reasonableindex system, which is applied for evaluating urbanization of a region andleading a reasonable development for its urbanization. Urbanization is notonly the urbanized process for population, but also for economicdevelopment and living style, so it is fnlly necessary to compare the urbanization level at all round.The majority studies for urbanization index system are subject to thequalitative analysis. The studies involved with the urbanization ofindividual provincial city or the urbanization levels for part of Chinesecities rather than all-round quantitative analysis for Chineseprefecture-level cities. This paper takes into account the all-round factorsto study the urbanization level in China. It adopts a compound indexsystem for urbanization to measure the urbanization with 17 indexes,which reflects the perspectives of economic development, living quality,social development quality, foundation establishment quality, entironmentquality and population development quality. With the index system forurbanization, this paper tries to supervise and evaluate the quantitativeenlargement and qualitative transition process of Chinese urbanizationwith clustering analysis methodology. The study applies a nonparametricKNN kernel density clustering method to analyze the panel data of 274prefecture-level cities for 2003 and 2004 and get mastery of not only thecommon feature of all cities but also the personality for individual city.The paper divides the urbanization level into six categories forcomparison analysis so as to explore the clustering feature of Chineseurbanization and the urbanization pattern of different cities, whichprovides important theoretical reference for management department andaccelerates the urbanization development with the pace of industrialization.