Synergistic effect of knowledge base and knowledge network on digital innovation in Chinese cities

This study, based on urban digital innovation patent data, employs the standard deviational ellipse method and a double machine learning model to systematically analyze the spatiotemporal evolution of digital innovation in Chinese cities and the multi-scale interactive effects between knowledge base and network structure. The results show that: 1) China’s digital innovation exhibits a “strong south-weak north, strong east-weak west” pattern, with the innovation center shifting northwestward, spatial coverage expanding, and a stable northeast-southwest dominant orientation; 2) Knowledge diversity, specialization, and related variety all positively contribute to innovation, indicating that both diverse knowledge integration and deep path development enhance innovation capacity; 3) Both betweenness and degree centrality significantly improve innovation level, with betweenness centrality playing a more prominent role, highlighting the importance of bridge nodes in knowledge flow; 4) Significant synergies exist between knowledge base and network structure, the interaction effect between related variety and network centrality is the strongest, indicating that within a context of related variety, network-based knowledge is more easily recombined and transformed; 5) There is notable urban heterogeneity in the interaction effects—cities with abundant knowledge and strong networks show positive synergies, medium-resource cities show no significant effect, and cities with weak resources or networks may experience negative effects, indicating potential efficiency losses due to knowledge concentration. This study enhances the understanding of knowledge element synergy in digital innovation and provides theoretical support for optimizing regional innovation policies.