Identification of Potential Cooperation Relationships Among Scientists.

In this article, the phenomenon of scientist cooperation in the scientist cooperation network is studied from the perspectives of information spread and link prediction. By mining the information in the scientist cooperation network, analyzing the cooperation has been generated and discovering potential cooperation opportunities. It helps to build a richer cooperation network with more content. Information spread can reflect the inner laws of network structure formation, and the link prediction method can retain the integrity of network information to the maximum extent. First, the real network is abstracted by analyzing its structure as well as node attributes into a simulated network. Second, the process of information spread in the cooperation network is simulated by improving the traditional SIS model. Some improvements are made to the link prediction algorithm for the impact brought to the network by information spread. Finally, the experimental results in the scientist cooperation network show that the hybrid weighted link prediction algorithm combining node attributes and spread factors can improve the accuracy of link prediction and provide suggestions for scientists to find partners. The comparative experiments on simulated and real networks not only validate the effectiveness of the propagation model in the scientist cooperation network, but also verify the accuracy of the hybrid weighted link prediction algorithm.