How Does Label Propagation Algorithm Work in Bipartite Networks?

Community detection in networks receives much attention recently. In this paper, we investigate a recently proposed community detection algorithm—label propagation algorithm (LPA), and propose a new algorithm to make it more suitable for bipartite networks. We demonstrate through five real world data sets that the new algorithm, while maintaining the accuracy and the near linear time complexity of the standard LPA, is more stable than the latter. More importantly, the new algorithm is readily parallelized, which implies that it is potential for real time community analysis of large-scale bipartite networks.