关键词: 全球创新网络, 跨国专利申请, 结构特征, 时间指数随机图模型（TERGM）

Abstract:     Integrating deeply into the global innovation network has become an important path for a country to acquire innovative breakthrough knowledge and enhance international competitiveness. Based on the data of international patent application from 1998 to 2020, this paper constructed the global innovation networks and analyzed their structural characteristics and evolution trend. Temporal Exponential Random Graph Model (TERGM) was used to deeply explore the influencing mechanism of the network from such two aspects as exogenous mechanism and endogenous mechanism. The study found that: (1) the global innovation network has become increasingly close, forming an unbalanced pattern concentrated in North America, Western Europe, the Oceania and East Asia. China′s status is significantly prominent, replacing Japan as the most influential network core country in Asia, second only to the United States. (2) The formation and evolution of the relationship of global innovation networks are affected by both exogenous and endogenous mechanisms. Countries with developed economies, high innovation capacity and large export scale are more likely to send patent applications, while large market countries with certain innovation capacity are more likely to become patent recipients. (3) Countries tend to form reciprocal patent application relationships. Under the effect of structure dependence and time dependence, countries are highly connected and have high transmissibility, and the patent application relationship chain continues to extend. This study has important practical significance for China to further improve its international patent applications and layout, deeply integrate into the global innovation network, enhance its technological autonomy, and cultivate new international competitive advantages for enterprises.

Key words: global innovation network, international patent application, structural characteristics, temporal exponential random graph model (TERGM)