Dynamic Detection of Academic Team Communities Based on Temporal Coauthor Network

As the status of team members and relationships between them change over time, the importance of membership nodes in the academic team also change, so that the academic team organizational structure evolves. The changes caused by the core members of the team led to the evolution of the team structure. Therefore, this paper presents a dynamic community discovery algorithm based on temporal coauthor network. By detecting the importance of nodes, the strength of relative edges, and its variation of persistence with time, the proposed algorithm implements creation, extension, shrink, division and disappearance operations on the communities in order to achieve the purpose of dynamic community discovery. In addition, for assessing the quality of community division, the paper proposes a method based on the interest similarity of Chinese key characters feature of academic teams. In experiment, a public document record dataset annodes in the academic team also change, so that the academic team organizational structure evolves. The changes caused by the core members of the team led to the evolution of the team structure. Therefore, this paper presents a dynamic community discovery algorithm based on temporal coauthor network. By detecting the importance of nodes, the strength of relative edges, and its variation of persistence with time, the proposed algorithm implements creation, extension, shrink, division and disappearance operations on the communities in order to achieve the purpose of dynamic community discovery. In addition, for assessing the quality of community division, the paper proposes a method based on the interest similarity of Chinese key characters feature of academic teams. In experiment, a public document record dataset and several syntheticd several synthetic dataset are used to verify the effectiveness our algorithm.