Abstract:    As an important part of national innovation, improving regional innovation performance is of great significance in enhancing the effectiveness of the national innovation system and supporting new breakthroughs in high-quality economic development. The intricate influence of network position on innovation performance has always been the focus of scholar′s attention, and previous studies have found that network position has a facilitating, inhibiting and nonlinear relationship on innovation performance. To bridge the gap between the theories, an attempt is made to deepen the understanding of the relationship between network position and differentiated knowledge base on regional innovation performance using the knowledge base view. For the network position theory, drawing on the knowledge base view literature that "heterogeneous knowledge is more valuable", we measured the scale of knowledge value in the network from the perspective of heterogeneous knowledge and incorporated the knowledge value into the research framework of network position and regional innovation performance. For the theory of differentiated knowledge base, we constructed indicators of "differentiated knowledge base" based on the network perspective and the view of knowledge bases and established the "relevant" relationship with external knowledge to enrich the research on relevant diversification in differentiated knowledge base.Based on Chinese patent data, the panel negative binomial regression method was used to analyze the non-linear effects and interactions of network position and differentiated knowledge base on regional innovation performance. The study found that: (1) structural hole has a "U-shaped" curve relationship with regional innovation performance, and the interaction effect with differentiated knowledge base is significant; (2) centrality has an "inverted Ushaped" curve relationship with regional innovation performance, and the interaction effect with differentiated knowledge base is not significant; and (3) differentiated knowledge base positively promote regional innovation performance. The significant difference between the interaction effects of network position and differentiated knowledge base on regional innovation performance suggests that the impact of the value of resources acquired through network position on regional innovation performance cannot be ignored.The study has offered several insights. First, when formulating regional innovation policies, the government should consider the network position between regions and its impact on innovation performance. Encouraging interregional firms and academic institutions to strengthen ties and moderately increase regional centrality in the network can enhance innovation. Second, the government should view the differentiated knowledge base as a key driver of innovation. By diversifying the industrial structure and creating open innovation platforms, regions can integrate diverse knowledge and resources, thus boosting innovation. Lastly, a comprehensive innovation performance evaluation mechanism should be established. Monitoring and optimizing the network position and knowledge base can lead to overall regional innovation improvement.

Key words: network position, differentiated knowledge base, regional innovation performance, knowledge-based view, interaction effect