Mining Key Scholars via Collapsed Core and Truss

Mining key scholars is an important task in various scenes such as experts finding, collaborator recommendation, etc. With the generalization of collaboration, the co-authorship network has extremely enlarged, which makes this task more difficult to be handled. In this work, we propose two algorithms to mine key scholars based on collapsed k-core/k-truss and closeness centrality. The proposed algorithms mine key scholars mainly from two perspectives, i.e., component segmentation calculation and auxiliary decision metric. The proposed algorithms are irrelevant to network connectivity, which performs with high efficiency. We employ closeness centrality to avoid the derivation of greedy strategy. By comparing with other baseline methods, we find that the two proposed algorithms outperform with higher efficiency and better effectiveness.