| Contents |
| Hidden geometries underlying real networks provide a simple and natural explanation for their large-scale structure and growth dynamics. These models assume a newtonian-like connection probability between nodes that encodes a trade-off between popularity and similarity, such that nodes closer in space --that is, more similar-- are more likely to interact and nodes with more connections --that is, more popular-- can reach further neighbors. Beyond their ability to simulate the observed topologies (including scale-free degree distributions, high levels of clustering, and self-similarity), these models enable a true cartography of real networks which can be embedded into a metric space using an optimization method. Applications to real systems range from metabolic networks to the Internet at the autonomous systems level. In economic systems, the gravity theory of trade flows can be reformulated to explain the observed topology of the world trade web in terms of a hidden metric space network model. Distances between two countries in the hidden geometry would correspond to an integrated measure that incorporates different distance attributes. We plan to map the world trade web such that the different countries are placed in relative positions according to not only their geographic location but also the actual aggregated barriers to international trade in the world. |
|