我国制造业结对集聚水平测度及其特征——机器学习方法的改进与应用

Measurement and Characteristic of Manufacturing Co-agglomeration in China:Improvement and Application of Machine Learning

结对集聚是产业空间集聚的重要现象之一。文章应用并改进机器学习方法,测度了中国城市的制造业结对集聚指数,从时序变化、空间格局方面分析结对集聚特征。研究表明:(1)制造业结对集聚指数整体上呈现出“U”型分布特征。(2)2013年劳动与劳动、劳动与资本、劳动与技术密集型产业显著结对集聚组合明显上升。(3)中国四大区域板块中,技术密集型产业的结对集聚指数低于劳动、资本密集型产业。(4)2013年城市产业空间网络呈现出一定的“核心—边缘”特征。最后,提出了加强产业链的整体规划与布局、进一步提升制造业集聚水平、引导城市制造业差异化发展等政策建议。

Co-agglomeration is an important feature of urban industrial spatial agglomeration. This article improves and applies machine learning methods to measure the manufacturing co-agglomeration index of 280 cities in China. The characteristic analysis shows that: 1) The manufacturing co-agglomeration exhibits a "U-shaped" distribution. 2) In 2013,there was a big increase of significant co-agglomeration combination of labor and labor, labor and capital, labor and technology intensive industries. 3) The co-agglomeration index of technology intensive industries in the four major sectors is lower than that of labor,capital intensive industries. 4) In 2013,the urban industrial spatial network showed a certain "core-edge" characteristic. Policy suggestions are put forward to strengthen the overall planning and layout of the industrial chain, further improve the level of manufacturing industry agglomeration, and guide the differentiated development of urban manufacturing industry.