基于无监督学习方法的收缩城市识别研究--以粤港澳大湾区9座地级市为例

当今世界,收缩城市作为一种全球性、地方性、复杂性、多维性的现象,逐渐引起学界和社会广泛关注。基于此,论文以粤港澳大湾区城市群中9座地级市为例,从经济、人口、空间地理、行政4个维度出发,通过2008-2017年各城市统计面板数据、灯光遥感数据DMSP/OLS的采集,初步建立了由44项潜在反映收缩城市特征的指标组成的识别体系。在此基础上,论文利用无监督学习中的K均值聚类算法与定量方法中的因子分析等方法划分了城市类别,并通过定性分析初步探究了其中收缩城市的形成原因。结果显示:根据论文最终构建的收缩城市综合识别体系对所研究城市的考察,肇庆、江门、惠州属于人口流失型收缩城市,具有低城市扩张水平、低城镇化水平、低第二、三产业就业水平等特点。此外,(1)深圳、(2)广州、(3)珠海、中山、东莞和佛山,分别被划分为(1)全面型扩张城市、(2)空间稳定型扩张城市和(3)稳定型城市。而自然条件局限、地区政策引导缺位、区域内基础设施不配套和人口年龄结构变动是该地区收缩城市出现的成因。

Research on Shrinking City Identification Based on Unsupervised Learning Method--A Case Study of 9 Prefecture-Level Cities in Guangdong-Hong Kong-Macao Greater Bay Area

In today’s world of urban expansion,some cities are actually shrinking.This phenomenon which is both global and local,complex and multi-dimensional,has attracted widespread attention from academic circles and society. This study takes nine prefecture-level cities in the Guangdong-Hong Kong-Macao as an example. Using the four dimensions of economy,population,spatial geography,and administration,the statistical panel data and light remote sensing data DMSP of each city for the period 2008-2017 were collected. This established an identification system consisting of 44 indicators of potential reaction-shrinking urban characteristics. On this basis,the k-means clustering algorithm was used for unsupervised learning and factor analysis in quantitative methods to classify urban categories,and qualitatively analyzed the reasons for the formation of the contracted cities. The results show that according to the urban comprehensive identification system constructed by this paper,Zhaoqing,Jiangmen and Huizhou are population loss-type contraction cities. This means they hadrelatively low urban expansion, urbanization levels, and industrial employment. In addition,Shenzhen was classified as an all-round expansioncity,Guangzhou was classified as a space-stable expansion city,Zhuhai,Zhongshan,Foshan,and Dongguan were classified as stablecities. Furthermore,the limitations of natural resources,the lack of regional policy guidance,the lack of infrastructure,and the age of the population were the causes for shrinking cities in the region.