西安市创新活动空间分布特征及影响因素研究——基于多尺度地理加权回归

Research on Spatial Distribution Characteristics and Infl uencing Factors of Innovation Activities in Xi’an:Based on Multi-Scale Geographically Weighted Regression

文章以专利申请数据表征创新活动,将其分为知识型、生产型创新活动两类,通过核密度分析法、空间自相关分析法、标准差椭圆、DBSCAN聚类等方法从总体和分类别两大层面分析和探究西安市创新活动的空间分布特征,得出其创新活动空间分布具有规律性分异、差异化集聚、选择性依赖三大特征的结论。在此基础上,利用多尺度地理加权回归模型,选取相关影响因素进一步分析创新活动的空间分异特征,最终得出创新主体驱动创新活动、环境配置吸引创新活动、知识借用激励创新活动的结论。

This paper uses patent application data to characterize innovation activities,divides them into two categories:knowledge-based and production-oriented innovation activities,analyzes and explores the spatial distribution characteristics of innovation activities in Xi’an at the overall and subcategory levels through kernel density analysis,spatial autocorrelation analysis,standard deviation ellipse,DBSCAN clustering and other methods,and concludes that the spatial distribution of innovation activities has three characteristics:regular diff erentiation,diff erentiated agglomeration and selective dependence.On this basis,the multi-scale geographically weighted regression model is used to further analyze the spatial diff erentiation characteristics of innovation activities by selecting relevant influencing factors,and finally concludes that innovation activities are driven by innovation subjects,attracted by environmental configuration,and motivated by knowledge borrowing.