Dynamic Community Updating Algorithm Using Secondary Influence Structure Similarity

Xin ZHANG,Yu ZHANG,You-juan YAO,Shan-zeng CHU,Xiao-guang LI
DOI: https://doi.org/10.3969/j.issn.1000-1220.2018.04.027
2018-01-01
Abstract:Current networks in various fields have the characteristic of huge scale,complex structure and dynamic change.Traditional community detection algorithms can not adapt to this kind of dynamic complex network very well.In order to solve the proplems of the detection dynamic changes,a static community detection algorithm-LBS is proposed. LBS is based on the secondary influence structure similarity and the related definitions,and further a dynamic community update algorithm-IU-LBS is proposed.LBS constructs chains by the connection preference of nodes,and then combines the chains into community structures.LBS can detect community with low time and high quality,and has advantages in the community incremental updating.On this basis,IU-LBS realize the incremental updating of the community on the dynamic complex network.IU-LBS checks the related nodes when the network changes,and uses the corresponding update strategy for different classes of nodes.Through experiments,compared with the classical algorithm on several real networks and analog networks,the results proved that IU-LBS has higher detection quality and better time efficiency.
What problem does this paper attempt to address?