Improving infectious disease predictions through the use of metapopulation SIR modeling and graph convolutional neural networks

Authors

  • Petr Kisselev Department of Mathematical Sciences, George Mason University, Fairfax, VA
  • Alonso G. Ogueda-Oliva Department of Mathematical Sciences, George Mason University, Fairfax, VA
  • Padmanabhan Seshaiyer Department of Mathematical Sciences, George Mason University, Fairfax, VA

Abstract

Graph convolutional neural networks have shown tremendous promise in addressing data-intensive challenges in recent years. In particular, some attempts have been made to improve predictions of SIR models by incorporating human mobility between metapopulations and using graph approaches to estimate corresponding hyperparameters. In [1], researchers have found that a hybrid GCN-SIR approach outperformed existing methodologies when used on the data collected on a precinct level in Japan. In our work, we extend this approach to data collected from the continental US, adjusting for the differing mobility patterns and varying policy responses. Extensions and generalizations of the metapopulation GCN-SIR learning framework are proposed.

 

[1] Cao, Q., Jiang, R., Yang, C., Fan, Z., Song, X., Shibasaki, R., “MepoGNN: Metapopulation epidemic forecasting with graph neural networks’’, Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pp. 453–468 (2022)

Published

2024-10-13

Issue

Section

College of Science: Department of Mathematical Sciences