MCEIL: An Improved Scoring Function for Overlapping Community Detection using Seed Expansion Methods

摘要

Community detection is one of the most well known problems in complex network analysis. In real-world networks, communities often overlap. Various approaches have been proposed in the literature to detect overlapping communities in networks. Local Expansion and optimization approaches have gained popularity due to their scalability and robustness. In a method based on local expansion, the seeding strategy and scoring function employed are crucial to the performance of the algorithm.In this paper, a scoring function called CEIL score is used with ground-truth seeds in local expansion and optimization algorithm. Using CEIL score has significantly improved performance of the algorithms with respect to evaluation metrics NMI and F1 score. However, CEIL has lower coverage than conductance. An extension to CEIL score, called MCEIL score is proposed. Using MCEIL score returns communities with coverage as high as conductance, and NMI and F1 scores higher than conductance on different kinds of datasets.Experiments on datasets of different types with different seeding strategies show that the improvements in NMI and F1 score obtained by MCEIL score are substantial.