LAUSR

Several research works had been carried out to discover suitable algorithms to quantify node centralities. Among the many existing centrality metrics, only few consider centrality at the sub-graph level or deal with structural hole capabilities of pivot nodes. Research has proven the importance of sub-graph information in distinguishing influential nodes. In this work, two centrality metrics are proposed to distinguish and rank nodes in complex networks. The first metric called Sub-graph Degree Information centrality is based on entropy quantification of a node’s sub-graph degree distribution to determine its influence. The second metric called Sub-graph Degree and Structural Hole centrality considers a node’s sub-graph degree distribution and its structural hole property. The two metrics are designed to efficiently support weighted and unweighted networks. Performance evaluations were done on five real world datasets and one artificial network. The proposed metrics were equally compared against some classic centrality metrics. The results show that the proposed metrics can accurately distinguish and rank nodes distinctly on complex networks. They can equally discover highly influential and spreader nodes capable of causing epidemic spread and maximum network damage.