Research on Interaction Tracking between Community Discovery and theme Evolution Based on DBLP Scientific Research Cooperation Network

摘要

To address the shortcomings of the evolution of the fragmentation theme and the evolution of the network structure itself in the cooperation network, this paper introduces the ideas and techniques of complex network research. It firstly discovers the whole time-slice network, and then introduces BLPA (Balanced Label Propagation Algorithm) to extract the mobile author nodes and the corresponding community theme and the mobile author node theme based on community discovery. A research method combining vertical theme distribution and horizontal theme evolution is designed to analyze the interaction evolution mechanism or law between theme evolution and community evolution. The experiment found that there are theme differences between different communities in the same time-slice, the mobility of co-workers tends to be dispersed, the mobile author nodes guide the theme change, and the theme evolution driven by community evolution is in line with the overall theme evolution trend. This method makes use of complex network related technologies to explore the intrinsic dynamic mechanism of community evolution, community theme evolution and evolution on DBLP datasets, which is realistically feasible.