ABCDPRGM (Attractor-Based Coevolving Dot Product Random Graph Model) is a dynamic latent space model designed to capture evolving network behaviors such as flocking or polarization. It extends the classic Random Dot Product Graph (RDPG) framework by introducing group-based attractors that guide the movement of node latent positions over time.
To reproduce the main synthetic experiments and figures from the paper, use the Jupyter notebook:
This notebook will:
- Generate synthetic dynamic networks based on the ABCDPRGM model
- Simulate polarizing or flocking behavior under different parameter settings
- Estimate the parameters
$(\beta_1, \beta_2, \beta_3, \beta_4)$ from the generated graphs - Output trajectory plots and other visualizations used in the paper
You can modify values such as:
- Number of nodes
- Number of time steps
- Initial Dirichlet parameters
- Influence coefficients (β)
No additional configuration files are required. All dependencies are standard Python scientific packages (see requirements.txt).
To reproduce the plots from the paper, use
To replicate the real-world data analysis presented in the paper, use the notebook:
This notebook performs the following:
- Loads and preprocesses ranked match data from Age of Empires IV
- Constructs time-series networks based on player interactions
-
Estimates model parameters
$(\beta_1, \beta_2, \beta_3, \beta_4)$ using the ABCDPRGM framework - Visualizes the evolution of latent positions to identify flocking or polarization behaviors
Note: The dataset is sourced from aoe4world.com/dumps. Please ensure you download the appropriate data corresponding to the time frame specified in the paper.
Disclaimer: The data is provided by aoe4world under Microsoft's "Game Content Usage Rules" using assets from Age of Empires IV, and it is not endorsed by or affiliated with Microsoft.