Preferential attachment: a multi-attribute growth process generating scale-free networks of different topologies

Abstract: This paper expands the degree-based consideration of the preferential attachment growth process and applies five different connectivity criteria (node degree, clustering coefficient, betweenness centrality, closeness centrality, and eigenvector centrality) to define the development of new links in the networks. Based on statistical inference, the analysis shows that all the available control attributes are capable generating SF networks, that the proposed generalized preferential attachment growth process produces networks of statistically different topologies, under different control-attributes, and that the betweenness centrality is the control-attribute generating networks of better topology. Overall, this paper introduces a multi-dimensional conceptualization of preferential attachment, which can motivate further research and can provide new tools for the modeling and interpretation of real-world networks that currently cannot be fully explained by the degree-driven BA models.