Enhancing Bibliometric Insights: A Novel Overlapping Clustering Framework for Citation Networks

The detection of communities within citation networks plays a pivotal role in understanding the structure and dynamics of scientific research. Traditional clustering methods often assume disjoint communities, overlooking the fact that publications frequently belong to multiple interconnected research domains. This paper introduces Overlapping kmp-clustering (OKMP), a scalable framework designed to uncover overlapping communities within large citation networks while preserving core-periphery structures. Unlike conventional methods, OKMP leverages a novel augmentation step to allow nodes to belong to multiple clusters, enabling richer interpretations of interdisciplinary influences. Using datasets from the exosome biology research community, we demonstrate the efficacy of OKMP compared to disjoint clustering methods and existing overlapping clustering techniques. The analysis highlights the relationship between citation metrics and cluster memberships, offering deeper insights into the network’s structural properties. Our findings emphasize the value of overlapping clusters in bibliometric research, paving the way for improved understanding of scientific collaboration and influence.