珠三角地区家具产业链空间聚类识别及空间熵分解

Spatial Cluster Identification and Spatial Entropy Decomposition of Furniture Industry Chain in the Pearl River Delta Region

文章基于珠三角地区全产业链家具企业POI数据,采用DBSCAN空间聚类算法识别并提取家具产业空间聚类集群,运用核密度估计、轮廓系数等方法剖析了家具产业空间聚类的总体特征、产业链结构特征和等级结构特征,并通过空间熵指数分解辨析了珠三角地区家具产业空间聚类的驱动力。结果表明:(1)珠三角地区家具产业主要聚类集群有7个,呈“多核心扇状”空间分布以及“一主三副”中心分布的等级结构特征,主中心位于顺德,副中心位于厚街、沙溪和香洲。(2)家具产业链上游企业空间聚类呈“二主一副”中心分布的等级结构,主中心为顺德和香洲,次中心为斗门;中游和下游企业空间聚类呈“一主一副”中心分布的等级结构,主中心均为顺德,次中心分别为禅城、厚街。(3)非空间因素更能解释珠三角地区总体家具产业的空间聚类,解释程度达80%,且随着空间距离的上升而逐渐增加。而空间因素主导家具产业链上、中、下游空间聚类,解释程度为60%~70%。(4)珠三角各地区家具产业空间聚类的形成与演进受空间与非空间因素的影响程度各不一样,广州、中山、江门等地家具产业空间聚类集群受空间因素的影响相比其他地区更突出。

Based on the POI data of furniture enterprises in the whole industrial chain in the Pearl River Delta(PRD)region,this paper uses the spatial clustering algorithm of DBSCAN to identify and extract the spatial clustering clusters in the furniture industry,analyzes the overall characteristics,industrial chain structure characteristics,and hierarchical structure characteristics of spatial clustering in the furniture industry by the means of the kernel density estimation and the silhouette coefficient,and explores the driving forces of spatial clustering of furniture industry in the PRD region through the decomposition of spatial quotient index.The results show that:1)There are 7 main clusters of furniture industry in the PRD region,which shows the"multi-core and fan-shaped"spatial distribution and the hierarchical structure characteristics of"one main center and three sub-centers",with the main center located in Shunde and the sub-centers located in Houjie,Shaxi and Xiangzhou.2)The upstream enterprises of the furniture industry chain show the spatial clustering distribution characteristics of"two main centers and one sub-center",taking Shunde and Xiangzhou as the main centers and Doumen as the sub-center.The midstream and downstream enterprises show the spatial clustering distribution characteristics of"one main center and one sub-center",taking Shunde as the main center and Chancheng and Houjie as the sub-centers.3)Non-spatial factors better explain the spatial clustering of the overall furniture industry in the PRD region,with the degree of explanation reaching 80%,and gradually increasing with the rise of spatial distance.On the contrary,spatial factors dominate the spatial clustering of the upper,middle and lower reaches of the furniture industry chain,with an explanation degree of about 60%to 70%.4)The formation and evolution of spatial clustering of the furniture industry in each region of the PRD is influenced by spatial and non-spatial factors to varying degrees,spatial factors influencing the spatial clustering of the furniture industry in Guangzhou,Zhongshan and Jiangmen more prominently than that in other regions.