数据驱动的产业集群辨识及空间关联性分析

Identification of Industrial Clusters with Data-driven Method and Spatial Correlation Analysis

利用网络爬虫与自然语言处理技术,通过获取网络数据开展对产业集群的研究。在特定产业集群区域采用定向网络爬虫抓取企业网页信息并进行解析,获取包括公司名称、供应产品、注册时间和地址等结构化数据;通过对文本信息的处理,利用聚类算法分析进行网络层次聚类,同时结合了地理空间数据,利用自下而上的网络分析方法映射到集群实体空间,对产业集群展开空间特性研究。基于聚类结果辨识出宁波地区注塑机集群7种具有不同产品结构的主体,结合核密度估计法与空间关联分析法,从时序与空间的角度讨论注塑机集群特性以及集群内部成员之间的关联性,为基于网络数据研究产业集群的结构提供了方法。

The research of industrial cluster is carried out by acquiring network data and using the technology of web crawler and natural language processing.In a specific cluster area,directional web crawler is used to obtain and analyze the enterprise information of web pages,including structured data of company name,supply products,registration time and address.Through the processing of text information,hierarchical clustering algorithm is used to analyze the industrial clusters.The research of the spatial characteristics of industrial clusters is conducted by using the bottom-up network analysis method to map the practical space with geospatial data.Seven DKM injection machine cluster with different product structures were identified based on the clustering results in Ningbo.Then combining kernel density estimation and spatial correlation analysis methods,the cluster characteristics of DKM injection machine and the correlation between members in the cluster were discussed from the perspective of time and space,which provides a method for researching the structure of industrial cluster based on network data.