Community detection in social networks using a novel algorithm without parameter

摘要

With regard to the expansion of networks like computer network and virtual social networks, many researchers have been drawn to the analysis of social networks. Community detection in social networks is one of the most important issues in this domain. In recent years, numerous algorithms have been introduced in order to detect communities. In the present study, by determining the number of clusters and specifying initial centres of clusters, a novel algorithm has been introduced. By defining the neighbourhood of nodes, the proposed algorithm selects the initial centres of clusters in a way that the centres are at the maximum distance from each other. This algorithm has been evaluated using the three known and commonly used algorithms, i.e., K-means, k-means++ and K-Rank-D over the Dolphin, Karate club and Soccer team datasets. The results show the superiority of the proposed algorithm over other algorithms.