Common neighbour structure and similarity intensity in complex networks

Hou, L. and Liu, K. (2017) Common neighbour structure and similarity intensity in complex networks. Physics Letters A, 381 (39). pp. 3377-3383. ISSN 0375-9601 doi: 10.1016/j.physleta.2017.08.050

Abstract/Summary

Complex systems as networks always exhibit strong regularities, implying underlying mechanisms governing their evolution. In addition to the degree preference, the similarity has been argued to be another driver for networks. Assuming a network is randomly organised without similarity preference, the present paper studies the expected number of common neighbours between vertices. A symmetrical similarity index is accordingly developed by removing such expected number from the observed common neighbours. The developed index can not only describe the similarities between vertices, but also the dissimilarities. We further apply the proposed index to measure of the influence of similarity on the wring patterns of networks. Fifteen empirical networks as well as artificial networks are examined in terms of similarity intensity and degree heterogeneity. Results on real networks indicate that, social networks are strongly governed by the similarity as well as the degree preference, while the biological networks and infrastructure networks show no apparent similarity governance. Particularly, classical network models, such as the Barab´asi-Albert model, the Erd¨os-R´enyi model and the Ring Lattice, cannot well describe the social networks in terms of the degree heterogeneity and similarity intensity. The findings may shed some light on the modelling and link prediction of di�erent classes of networks.

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/76181
Identification Number/DOI	10.1016/j.physleta.2017.08.050
Refereed	Yes
Divisions	Henley Business School > Digitalisation, Marketing and Entrepreneurship
Publisher	Elsevier
Download/View statistics	View download statistics for this item