Human dynamics in mobile social networks: A study of inter-node relationships

To better understand human mobility patterns and interactions, it is important to go beyond the behavior of the individual level and mine the implicit correlation from the anonymized traces. We empirically studied some statistical properties of direct contact traces that recorded contacts between short range wireless communication devices. To describe internode relationships, we introduced a distance metric, i.e. Closeness Matrix. Closeness Matrix was inspired by the fact that, in social systems, technological systems, and biological systems, closely connected individuals tended to share common adjacent members. The results show that (1) mobility patterns of the individual level do not agree with that of population level since people have different personalities and behaviors; (2) closely connected individuals have similar power law exponents. Such understanding may promote more efficient behavior-aware protocols and applications.