A Community Detection Method Based on Local Optimization in Social Networks

Abstract

Social networks are a kind of communal structure composed of many nodes according to different kinds of relationships. Community detection can help people understand the topology of a network and identify meaningful clusters. A large variety of community detection algorithms have been proposed. However, many of these algorithms are suitable for unweighted networks since they ignore the social properties of the links between nodes. Meanwhile, their discovery results may lead to the emergence of huge communities. Therefore, a weighted method and local community stability for social networks based on LFM is proposed in this article. This WSLFM algorithm uses a weighted method based on the social attributes of the link and the degree of common neighboring nodes to update the fitness function. It also introduces the concept of stability to control the expansion of local communities. The proposed WSLFM algorithm was tested and proved to be effective on a synthetic network and a real-world network for finding smaller and more meaningful clusters.