Abstract:We introduce Polaris, a network null model for colored multi-graphs that preserves the Joint Color Matrix. Polaris is specifically designed for studying network polarization, where vertices belong to a side in a debate or a partisan group, represented by a vertex color, and relations have different strengths, represented by an integer-valued edge multiplicity. The key feature of Polaris is preserving the Joint Color Matrix (JCM) of the multigraph, which specifies the number of edges connecting vertices of any two given colors. The JCM is the basic property that determines color assortativity, a fundamental aspect in studying homophily and segregation in polarized networks. By using Polaris, network scientists can test whether a phenomenon is entirely explained by the JCM of the observed network or whether other phenomena might be at play. Technically, our null model is an extension of the configuration model: an ensemble of colored multigraphs characterized by the same degree sequence and the same JCM. To sample from this ensemble, we develop a suite of Markov Chain Monte Carlo algorithms, collectively named Polaris-*. It includes Polaris-B, an adaptation of a generic Metropolis-Hastings algorithm, and Polaris-C, a faster, specialized algorithm with higher acceptance probabilities. This new null model and the associated algorithms provide a more nuanced toolset for examining polarization in social networks, thus enabling statistically sound conclusions.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to propose a new network null model, which is specifically used to study network polarization phenomena. Specifically, the author introduced Polaris, a network null model for colored multigraphs that preserves the Joint Color Matrix (JCM). The JCM specifies the number of edges connecting vertices of any two colors, which is a fundamental property that determines color assortativity and is crucial for studying homogeneity and segregation phenomena.
### Problem Background
In network science, when studying polarization phenomena, network representations are usually used, and a rich set of network and graph - theory tools are used to define properties and calculate relevant metrics. However, these metrics are only meaningful if they are not statistical noise. For this reason, null - hypothesis models are used to assess statistical significance. But existing network null models are too simple, usually only preserving basic graph - structure features such as density or degree sequence, while ignoring the interactions between opinions or communities that describe polarization phenomena.
### Contribution of Polaris
Polaris proposes a new network null model, which is particularly suitable for studying network polarization. This model is a statistical ensemble consisting of colored multigraphs, where each vertex has a color (i.e., a single label) and edges can appear multiple times. Polaris preserves two key properties:
1. **Degree sequence**: The degree of each vertex.
2. **Joint Color Matrix (JCM)**: Specifies the number of edges connecting vertices of any two colors.
By preserving the JCM, Polaris can better capture color assortativity in the network, which is very important for studying homogeneity and segregation phenomena.
### Technical Implementation
To sample from this ensemble, the author developed a series of Markov Chain Monte Carlo (MCMC) algorithms, collectively known as Polaris - *. These include:
- **Polaris - B**: An adapted version based on the general Metropolis - Hastings algorithm.
- **Polaris - C**: A faster and more specialized algorithm with a higher acceptance probability.
These new algorithms provide a more detailed set of tools for examining polarization phenomena in social networks, leading to statistically reliable conclusions.
### Summary
In general, this paper aims to solve the problem that existing models cannot effectively capture network polarization phenomena by proposing Polaris, a new network null model. By preserving the Joint Color Matrix, Polaris can more accurately simulate and analyze homogeneity and segregation phenomena in polarized networks, providing a more powerful tool for network scientists.