Latent space generative model for bipartite networks

TL;DR

This paper introduces a novel latent space generative model for bipartite networks, extending hyperbolic growth models to better capture structural properties like degree distributions and small cycles, enhancing understanding of bipartite network formation.

Contribution

It presents the first latent space generative model for bipartite networks based on hyperbolic geometry, improving modeling of their structural features.

Findings

Model reproduces bipartite degree distributions

Captures small cycles in bipartite networks

Enables assessment of projected network properties

Abstract

Generative network models are extremely useful for understanding the mechanisms that operate in network formation and are widely used across several areas of knowledge. However, when it comes to bipartite networks -- a class of network frequently encountered in social systems -- generative models are practically non-existent. Here, we propose a latent space generative model for bipartite networks growing in a hyperbolic plan. It is an extension of a model previously proposed for one-mode networks, based on a maximum entropy approach. We show that, by reproducing bipartite structural properties, such as degree distributions and small cycles, bipartite networks can be better modelled and one-mode projected network properties can be naturally assessed.