基于多尺度核心集群方法的粤港澳大湾区R&D集群识别  

Identification of R&D Clusters in the Guangdong-Hong Kong-Macao Greater Bay Area Based on Multiscale Core Cluster Approach

作　　者：朱丰毅 韩兆洲[1,2] 桂文林 ZHU Feng-yi;HAN Zhao-zhou;GUI Wen-lin(College of Economics,Jinan University,Guangzhou 510632,China;Guangzhou Huashang College,Guangzhou 511300,China)

机构地区：[1]暨南大学经济学院,广东广州510632 [2]广州华商学院,广东广州511300

出　　处：《数理统计与管理》2024年第5期811-829,共19页Journal of Applied Statistics and Management

基　　金：国家社会科学基金重点项目(19ATJ004);暨南大学高性能计算公共服务平台支持。

摘　　要：本文引入标记加权局部Ripley's K函数,改进了多尺度核心集群方法,能更精确地识别集群核心点。在实证部分,本文首先采用Web API对2010年-2019年粤港澳大湾区(大湾区)的发明专利数据获得真实R&D点模式,基于工业GDP和企业POI构造反事实点模式。接着,利用改进的方法进行R&D集群识别,在多个空间尺度下分析集群的空间特征和演化过程。本文发现:(1)大湾区的R&D集聚度显著高于生产活动且逐年提高。(2)R&D集聚度存在行业差异,部分行业的R&D集聚度显著高于全样本。(3)集群沿东西轴线分布,广州、深圳是两个创新极点,东岸轴线上形成了“广州-东莞-深圳-香港”集群。(4)东西轴线R&D发展不平衡。东岸集群规模大、活动更集中,而西岸集群规模较小且核心点较分散。(5)集群空间分布的行业差异较小,城市间R&D产业分工不明确。This paper introduced the mark-weighted local Ripley's K-function to adapt the multiscale core-cluster approach,which can identify the cluster core points more precisely.In the empirical part,firstly,we obtained the real R&D point patterns by using Web API to locate invention patents in the Guangdong-Hong Kong-Macao Greater Bay Area(Greater Bay Area)from 2010 to 2019,and counterfac-tual point patterns were constructed based on the industrial GDP and firm POI.Then,we employed the adapted method to identify R&D clusters and analyzed the spatial characteristics and evolution process of the clusters at multiple spatial scales.We found that:(1)Degree of R&D agglomeration in Greater Bay Area was significantly higher than that of production activities and the agglomeration degree in-creased by the year.(2)The R&D agglomeration degree varied with industries.Some industries were significantly more concentrated than the full sample.(3)Clusters were distributed along the east and west axis.Guangzhou and Shenzhen were the two innovation poles.On the axis of the East Bank of the Pearl River Estuary,the“Guangzhou-Dongguan-Shenzhen-Hong Kong”cluster had formed.(4)R&D on the west and east axis developed unbalanced.The east bank clusters were large and activities were distributed more concentrated,while the west bank clusters were small in scale and their core points were more scattered.(5)The distribution of clusters had almost no industry differences,and the industrial division of R&D among cities was not clear.

关 键 词：标记加权局部Ripley's K函数 多尺度核心集群方法 R&D集群识别 发明专利 

分 类 号：F061.5[经济管理—政治经济学] O212[理学—概率论与数理统计]