Libraly Journal

Libraly Journal ›› 2023, Vol. 42 ›› Issue (389): 103-112.

Previous Articles     Next Articles

Research on Technology Opportunity Discovery Based onLink Prediction: Taking Patent Data in the Field of PrivacyProtection as an Example

Xu Xin1, 2, Li Qian1, Yi Chunbo1, Fan Yaxin (1 Faculty of Economics and management, East China Normal University; 2 Social Survey and Data Center, East China Normal University)   

  • Online:2023-09-15 Published:2023-09-21
  • About author:Xu Xin1, 2, Li Qian1, Yi Chunbo1, Fan Yaxin1 (1 Faculty of Economics and management, East China Normal University; 2 Social Survey and Data Center, East China Normal University)

Abstract:

In the era of scientific and technological competition, the identification and mining oftechnological opportunities lay an important foundation for the formulation of scientific and technologicalR&D guidelines and policies. A quantitative method based on link prediction is proposed to transformthe task of technological opportunity discovery into a binary problem of whether there can be cooccurrencelinks between technological elements, so as to provide detailed data assistance and supportfor technological opportunity identification. Obtaining patent data in the field of privacy protection,constructing IPC co-occurrence network, training GNN model GraphSAGE, and after obtaining thelink prediction score, the authors identify the technology opportunities for sustainable development,breakthrough and innovation and the technology opportunities for transformation from the perspectivesof enhanced, newborn and declining links respectively. Finally, they analyze and put forward suggestionsbased on the identified technology opportunities.