Research On Community Detection Algorithm Based On Label Propagation

With the development of society,the research on complex network community detection algorithm is gradually deepening.Community detection algorithm has great value in the field of recommendation,information dissemination,precision marketing and so on.The label propagation algorithm is favored by researchers because of its simplicity and high efficiency.However,label propagation community detection algorithm has the problems of poor stability and low accuracy.To solve these problems,this paper proposes a community detection algorithm based on label propagation in static community and dynamic community.Firstly,aiming at the lack of stability of label propagation algorithm,a label propagation overlapping community detection algorithm OCKELP is proposed,which combines label entropy and k-shell.In this algorithm,K-shell algorithm is used to initialize and label the node with the largest K value;label entropy is used to update the label from small to large order and asynchronous update strategy,and comprehensive influence is introduced when selecting the label,combining the community level information and local information;the termination condition of OCKELP algorithm is described,and the time complexity of the algorithm is analyzed.Secondly,aiming at the stability of label propagation algorithm and the low accuracy of incremental dynamic community detection,an incremental dynamic community detection algorithm IDCELP based on label propagation is proposed.The community structure at the initial time is obtained by OCKELP algorithm;the incremental information of adjacent time snapshots is used to find out the differentiated nodes and their neighbors to join the active node list,and the KELP process is used to update the label of the active node list until the termination condition is satisfied,the algorithm stops.The time complexity of the algorithm is analyzed.Finally,the feasibility and stability of OCKELP algorithm are verified by experiments on real network and artificial network.Compared with other algorithms,OCKELP algorithm can get better community structure.Experiments on real dynamic network and artificial dynamic network show that IDCELP algorithm has good stability,makes use of historical community structure,improves time efficiency to a certain extent,and obtains good community detection results.