Label propagation community detection algorithm based on density peak optimization
Yan Ma, Chen Guoqiang · 2021 17th International Conference on Computational Intelligence and Security (CIS) · 2021
Label propagation algorithm is one of the popular community detection algorithms in recent years. The advantages of the community detection algorithm based on label propagation are that the algorithm logic is simple, compared with the modularity optimization algorithm, the convergence speed is very fast, the whole clustering process does not need any optimization function, and there is no need to specify the number of communities in the complex network before initialization. However, this algorithm has problems such as unstable partitioning results and strong randomness. In order to solve these problems, this paper proposes a semi-supervised label propagation community detection algorithm based on density peak. The proposed algorithm first introduces the density peak to discover the cluster centers, determines the prototype of community, fixes the number of communities and cluster centers in complex network, and then uses label propagation algorithm to detect communities, which improves the accuracy and robustness of community discovery, reduces the number of iterations, and accelerates the formation of the communities.