| Summary: | This paper introduces dependency centrality, a node-level measure of structural leadership in bipartite networks. The measure builds on Zhou et al.'s (2007) flow-based method to transform bipartite data and captures additional information from the second mode that existing measures of centrality typically exclude. Three previously published bipartite networks serve as test cases to demonstrate the extent of correlation among node-level centrality rankings derived from dependency centrality and those derived from canonical centrality measures: degree, closeness, betweenness, and eigenvector. Ultimately, dependency centrality appears to offer a novel means to measure importance in bipartite networks depicting social interactions.
|
|---|