Prox is a stochatic method designed to study the structure of real-world graphs (i.e. large-scale graphs that model real data).
Particles wander randomly over the studied graph from vertex to vertex by following its edges. The structural properties of the graph are highlighted by the dynamics of the paths borrowed by the particles.
When the graph is built from linguistic material (corpora and dictionaries), Prox helps define a lexical metrology (distance, classification, cartography, browsing).
The Prox method comes with an online visualization interface.
Overseer / ContactBruno Gaume
Prox interface: snapshots