Edge Label Propagation Algorithm Based on Node Influence

Abstract

Community discovery is an important tool for understanding complex networks in depth. For real-world complex networks, overlap of communities is essential. In order to discover overlapping communities, this paper presents an overlapping community detection method (ELPANI) using node influence and side label propagation. Since the birth of LPA, a wide range of attention has been paid, but its results are unstable due to the randomness of its label renewal order. To solve these problems, we use the node's influence to initialize the edge label, and in the edge label update process, we use the degree of edge sorting to avoid these random factors. In order to retain multiple communities, we have retained multiple tabs on the side. Finally, we restore the completed edge label to the node.