城市科技创新效率与网络结构特征——对国家级创新型城市的实证分析

以2003—2017年中国75个创新型城市面板数据为样本，首先，运用SBM超效率模型测度各城市科技创新效率；其次，基于修正引力模型测度城市间创新联系强度，运用社会网络分析法，从人才流动、资本流动和制度学习3个方面对创新型城市的网络结构及演变特征进行测度和分析；最后，利用考虑城市地理位置的地理加权回归（GWR）模型分析网络结构特征对科技创新效率的影响机制。结果发现：①科研基础设施和人才流动中心度对城市科技创新效率提升具有促进作用，而制度学习中心度、人才教育和产业结构为负向影响；②随着时间推移，城市人才流动中心度对科技创新效率提升具有积极促进作用且趋于强化，资本流动中心度从具有微弱的积极作用趋向于在不同城市呈现出激励和阻碍两种效应，制度学习中心度的负面影响亦趋于强化；③各城市创新网络结构对科技创新效率的影响在测度期前期差别较小，而在后期呈现出越发显著的空间差异性。

Urban Science and Technology Innovation Efficiency and Network Structure Characteristics:an Empirical Analysis Based on National Innovative Cities

Using the panel data of 75 innovative cities in China from 2003 to 2017,the Super-SBM model is used to measure the technological innovation efficiency of each city.Then,based on the strength of the innovative connection between cities obtained by the modified gravity model,the social network analysis method is used to measure and analyze the network structure and evolution characteristics of innovative cities from the three aspects of talent flow,capital flow,and institutional learning.Finally,the Geographic Weighted Regression (GWR) model considering the geographical location of cities is used to analyze the influence mechanism of network structure characteristics on the efficiency of technological innovation.The results show that：①The scientific research infrastructure and the centrality of talent flow can promote the improvement of urban scientific and technological innovation efficiency,but the impact of institutional learning centrality,talent education and industrial structure is negative; ②Over time,the centrality of talent flow in cities has a positive effect on the improvement of scientific and technological innovation efficiency and tends to strengthen.The centrality of capital flow has a weak positive effect and tends to show two effects: incentive and hindrance in different cities.The negative impact of institutional learning centrality also tends to strengthen;③The difference in the impact of the innovation network structure on technological innovation efficiency in each city is relatively small in the early part of the measurement period,and it shows more significant spatial differences in the later stage.