G-CaNDEM: A Causal Graph-Based Learning Model for Epidemic Interventions on Human Mobility Networks

Coaquira L.J.H.; Valdivia A.M.C. · Proceedings - International Conference of the Chilean Computer Science Society, SCCC (2025) · DOI: 10.1109/sccc67219.2025.11420403

Resumen

Human mobility plays a key role in the spread of infectious diseases, yet modeling epidemic dynamics is challenged by incomplete and heterogeneous data. We present G-CaNDEM, a causal graph-based model that combines Bayesian imputation, Gaussian process interpolation, and spatiotemporal graph neural networks to simulate disease propagation under different interventions. Applied in a real-world case study, G-CaNDEM outperforms baseline models, reducing mean absolute error (MAE) and mean squared error (MSE) by up to 37%. Counterfactual simulations indicate that targeted mobility restrictions on urban hubs can reduce epidemic spread by 8-12%. The framework also provides interactive visualizations of transmission routes and hubs, supporting public health planning. G-CaNDEM offers a robust approach for modeling epidemics on complex mobility networks with incomplete data.

Instituciones: Universidad Nacional de San Agustín de Arequipa