Visualizing and exploring modular networks based on a probabilistic model

TL;DR

This paper introduces a probabilistic model-based visualization method for modular networks, using local attributes to reveal structure and facilitate analysis in a low-dimensional space.

Contribution

It presents a novel approach to analyze network modularity through fitted probabilistic models and attribute-based visualization.

Findings

Attributes effectively visualize network modules.

Attribute space improves network analysis.

Model fitting captures network structure.

Abstract

We propose a method to investigate modular structure in networks based on fitted probabilistic model, where the connection probability between nodes is related to a set of introduced local attributes. The attributes, as parameters of the empirical model, can be estimated by maximizing the likelihood function of the observed network. We demonstrate that the distribution of attributes provides an informative visulization of modular networks on low-dimensional space, and suggest the attribute space can be served as a better platform for further network analysis.