S1xS1-model-for-Bipartite-Networks

Bipartite networks possess several properties of interest that latent geometry frameworks for standard networks are not able to capture. In this work, we first briefly introduce the features commonly observed in real-world bipartite graphs (power-law degree distributions, high number of common neighbours and bipartite clustering coefficients) and their corresponding mathematical definitions. We then review a powerful latent-geometry framework specific to bipartite networks and proceed to show, through numerical simulations, that this model maintains several of the aforementioned properties and give original insights on its limitations. Finally, we propose a novel Montecarlo-based approach for the inference of the hidden parameters underlying the latent geometry