Abstract:Industry-university collaboration（IUC）has become an important element for the innovation-driven development strategy and transformation of scientific and technological achievements in China. With the continuous development of IUC, more and more innovative enterprises and research universities join in the technology collaboration through resource share and complement of each other’s advantages, which form the IUC networks with multiplex actors,abundant relationships and complex structure. A part of present literature on IUC network focuses on the descriptive analysis of collaboration mode and identification of the key actors, and the other literature explores its effect on innovation outcomes. There is a lack of discussion on the dynamic mechanism of IUC network. Therefore, combining with network system dynamics,this paper explores the formation mechanism of IUC network from the technical, social and geographical dimensions.This paper takes IUC of Chinese artificial intelligence technology as a case. We measure the key indicators and construct IUC network based on the patent data of Chinese artificial intelligence technology. Specifically, this paper uses the information of patentee and international patent classification from 2000 to 2019 to measure the independent variables:three technical factors including technological diversity, technological value and technological similarity, social factors represented by structural embeddedness, geographical factors represented by geographical proximity. The IUC network was constructed using the patentee information from 2020 to2021 as the dependent variable of this study. Finally, the Exponential Random Graph Model（ERGM）is adopted to analyze the influence of above mentioned endogenous and exogenous factors on the evolution of IUC network. The analysis tool is software Pnet.The empirical results show that diverse technology base and similar technology background will promote universities and enterprises to join the IUC network. The results also indicate that universities with higher technological value are more likely to be selected as partners by enterprises. In addition, the IUC network shows an inverse core-marginal trend, i.e. the universities and enterprises will choose a small number of partners for in-depth collaboration. Finally, the positive results of geographical proximity show that the IUC is more likely to occur within the province area. Our research results enrich the theories related to the evolution of the IUC network and provide ideas and scientific basis for the government to guide and promote the IUC in an orderly manner.

Mining of Industry-University-Research Cooperation Relationship Based on Multi-layer Heterogeneous Network Link Prediction

A Multi-Channel Multi-Tower GNN Model for Job Transfer Prediction Based on Academic Social Network.

Partner Selection in China Interorganizational Patent Cooperation Network Based on Link Prediction Approaches

Patent collaborative filtering recommendation approach based on patent similarity

Multi-Tasking Link Prediction On Coupled Networks Via The Factor Graph Model

Effective Model Integration Algorithm for Improving Link and Sign Prediction in Complex Networks

Link Prediction for Statistical Collaboration Networks Incorporating Institutes and Research Interests

From technology opportunities to solutions generation via patent analysis: Application of machine learning-based link prediction

A Community Detection and Graph-Neural-Network-Based Link Prediction Approach for Scientific Literature

Heterogeneous Graph based Deep Learning for Biomedical Network Link Prediction

A link prediction method for heterogeneous networks based on BP neural network

Link prediction in heterogeneous information networks: An improved deep graph convolution approach

A Community Detection and Graph Neural Network Based Link Prediction Approach for Scientific Literature

Formation Mechanism of Industry—University Collaboration Net?work:An Empirical Study of Chinese AI Technology

Choosing the right collaboration partner for innovation: a framework based on topic analysis and link prediction

A Bi-Scale Method Of Link Prediction

Enhancement to community-based multi-relational link prediction using co-occurrence probability feature

PatentMiner: Patent Vacancy Mining via Context-enhanced and Knowledge-guided Graph Attention

Research on the Patent Innovation Performance of University–Industry Collaboration Based on Complex Network Analysis

How Do Heterogeneous Networks Affect a Firm's Innovation Performance? A Research Analysis Based on Clustering and Classification