Traditional community identification is often based on static network, ignoring the social network variability, and has a great limitation. With the deepening of node embedding theory, it is gradually applied to the field of dynamic community, which provides convenience for dynamic community identification research, effectively reduces the distance between nodes in the structure of dynamic community and makes the composition of nodes become relatively simple. Because the node system is very complicated in the process of dynamic community identification, the error of dynamic community identification is large and even the identification is invalid. In order to better identify the dynamic community, this paper explains the concept of node embedding and illustrates the advantages of node embedding for dynamic community recognition based on node embedding technology. On this basis, using spectrum clustering algorithm, and carries on the improvement introduced an incremental dynamic community recognition algorithm, is used to identify the relevant experimental study on the dynamic community and obtained the dynamic community recognition of the two methods: one is based on social network topology delta identification method and the second is a dynamic community recognition method based on node force, which not only broke through the limitations of traditional static community research, and improve the reliability of the dynamic community identification, accuracy and flexibility, to a certain extent, the relevant dynamic community recognition of the future research provides a useful guide.
CITATION STYLE
Zhang, X., Zhang, J., & Yang, J. (2020). Dynamic Community Recognition Algorithm Based on Node Embedding and Linear Clustering. In Lecture Notes in Electrical Engineering (Vol. 675, pp. 829–837). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-15-5959-4_102
Mendeley helps you to discover research relevant for your work.