Unfolding accessibility provides a macroscopic approach to temporal networks.
Hartmut H K Lentz1, Thomas Selhorst2, Igor M Sokolov3
1Institute for Physics, Humboldt-University of Berlin, Newtonstrasse 15, 12489 Berlin, Germany and Institute of Epidemiology, Friedrich-Loeffler-Institute, Seestrasse 55, 16868 Wusterhausen, Germany.
Physical Review Letters
|August 29, 2014
Summary
We introduce a method to analyze temporal networks by creating accessibility graphs. This approach reveals shortest path durations and time scales, applicable to social, trade, and contact networks.
Area of Science:
- Network Science
- Temporal Networks
- Graph Theory
Background:
- Traditional accessibility graphs represent static networks, linking nodes if a path exists.
- Temporal networks, where connections change over time, require new analytical methods.
- Static representations of temporal networks can lose crucial dynamic information.
Purpose of the Study:
- To generalize the concept of accessibility graphs to temporal networks.
- To develop methods for analyzing temporal network dynamics using accessibility graphs.
- To introduce a metric (causal fidelity) for evaluating static network representations.
Main Methods:
- Constructing temporal accessibility graphs by progressively adding paths of increasing length (unfolding).
- Analyzing the distribution of shortest path durations within temporal networks.
- Calculating characteristic time scales inherent in temporal network dynamics.
- Defining and applying causal fidelity to assess the accuracy of static network models.
Main Results:
- The unfolding method provides insights into shortest path duration distributions.
- Characteristic time scales of temporal networks can be effectively identified.
- Causal fidelity quantifies the quality of static representations for temporal networks.
Conclusions:
- The proposed method successfully extends accessibility graphs to temporal networks.
- This approach offers valuable information on network dynamics, path durations, and time scales.
- Demonstrated applicability across diverse real-world temporal networks (social, trade, sexual contacts).
