Modeling network growth with directional attachment and communities



Aiming to construct a more precise probabilistic model for the short-term growth process of real-world growing networks such as the Web, we propose a new network growth model and its learning algorithm. Unlike the conventional network growth models, we incorporate directional attachment and community structure for this purpose. We first show that the proposed model also exhibits a degree distribution with a power-law tail. In experiments using real Web data, we demonstrate that prediction of the dynamics of a given growing network can be improved by incorporating directional attachment and community structure. In experiments using synthetic data, we also demonstrate that prediction performance can definitely be improved by incorporating community structure. © 2004 Wiley Periodicals, Inc. Syst Comp Jpn, 35(8): 1–11, 2004; Published online in Wiley InterScience ( DOI 10.1002/scj.10706