Centrality prediction in dynamic human contact networks
Computer Networks: The International Journal of Computer and Telecommunications Networking
Centrality and mode detection in dynamic contact graphs; a joint diagonalisation approach
Proceedings of the 2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining
| Hi-index | 0.00 |
The measurement and the analysis of the temporal patterns arising in human networks is of fundamental importance to many application domains including targeted advertising, opportunistic routing, resource provisioning (e.g., bandwidth allocation in infrastructured wireless networks) and, more in general, modeling of human social behavior. In this paper we present a novel and exhaustive study of the temporal dynamics of human networks and apply it to different sets of wireless network traces. We consider networks of contacts among users (i.e., peer-to-peer opportunistic networks). We show that we are able to quantify how the amount of information associated to the process evolves over time by using techniques based on time series analysis. We also demonstrate how regular patterns appear only at certain time scales: network dynamics appears nonstationary, in the sense that its statistical description is different at various time scales. These results provide a new methodology to accurately and quantitatively investigate the temporal properties of any type of human interactions and open new directions towards a better understanding of the regular nature of human social behavior.