Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread
The spread of an epidemic over a population is influenced by a multitude of factors having both spatial and temporal nature, which are hard to completely capture using first principle methods. This paper concerns regional forecasting of SARS-Cov-2 infections 1 week ahead using machine learning. We e...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-12-01
|
| Series: | Frontiers in Physics |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fphy.2023.1277052/full |
| _version_ | 1797391235939303424 |
|---|---|
| author | V. Maxime Croft Senna C. J. L. van Iersel Cosimo Della Santina Cosimo Della Santina |
| author_facet | V. Maxime Croft Senna C. J. L. van Iersel Cosimo Della Santina Cosimo Della Santina |
| author_sort | V. Maxime Croft |
| collection | DOAJ |
| description | The spread of an epidemic over a population is influenced by a multitude of factors having both spatial and temporal nature, which are hard to completely capture using first principle methods. This paper concerns regional forecasting of SARS-Cov-2 infections 1 week ahead using machine learning. We especially focus on the Dutch case study for which we develop a municipality-level COVID-19 dataset. We propose to use a novel spatiotemporal graph neural network architecture to perform the predictions. The developed model captures the spread of infectious diseases within municipalities over time using Gated Recurrent Units and the spatial interactions between municipalities using GATv2 layers. To the best of our knowledge, this model is the first to incorporate sewage data, the stringency index, and commuting information into GNN-based infection prediction. In experiments on the developed real-world dataset, we demonstrate that the model outperforms simple baselines and purely spatial or temporal models for the COVID-19 wild type, alpha, and delta variants. More specifically, we obtain an average R2 of 0.795 for forecasting infections and of 0.899 for predicting the associated trend of these variants. |
| first_indexed | 2024-03-08T23:29:01Z |
| format | Article |
| id | doaj.art-932628fcaaed40848863f18bb406ff37 |
| institution | Directory Open Access Journal |
| issn | 2296-424X |
| language | English |
| last_indexed | 2024-03-08T23:29:01Z |
| publishDate | 2023-12-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Physics |
| spelling | doaj.art-932628fcaaed40848863f18bb406ff372023-12-14T16:21:38ZengFrontiers Media S.A.Frontiers in Physics2296-424X2023-12-011110.3389/fphy.2023.12770521277052Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spreadV. Maxime Croft0Senna C. J. L. van Iersel1Cosimo Della Santina2Cosimo Della Santina3Delft University of Technology, Delft, NetherlandsNational Institute for Public Health and the Environment (Netherlands), Bilthoven, NetherlandsDelft University of Technology, Delft, NetherlandsInstitute of Robotics and Mechatronics, DLR, Oberpfaffenhofen, GermanyThe spread of an epidemic over a population is influenced by a multitude of factors having both spatial and temporal nature, which are hard to completely capture using first principle methods. This paper concerns regional forecasting of SARS-Cov-2 infections 1 week ahead using machine learning. We especially focus on the Dutch case study for which we develop a municipality-level COVID-19 dataset. We propose to use a novel spatiotemporal graph neural network architecture to perform the predictions. The developed model captures the spread of infectious diseases within municipalities over time using Gated Recurrent Units and the spatial interactions between municipalities using GATv2 layers. To the best of our knowledge, this model is the first to incorporate sewage data, the stringency index, and commuting information into GNN-based infection prediction. In experiments on the developed real-world dataset, we demonstrate that the model outperforms simple baselines and purely spatial or temporal models for the COVID-19 wild type, alpha, and delta variants. More specifically, we obtain an average R2 of 0.795 for forecasting infections and of 0.899 for predicting the associated trend of these variants.https://www.frontiersin.org/articles/10.3389/fphy.2023.1277052/fullepidemic predictiondeep learningspatio-temporal graph neural networksreal world evidenceCOVID-19 |
| spellingShingle | V. Maxime Croft Senna C. J. L. van Iersel Cosimo Della Santina Cosimo Della Santina Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread Frontiers in Physics epidemic prediction deep learning spatio-temporal graph neural networks real world evidence COVID-19 |
| title | Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread |
| title_full | Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread |
| title_fullStr | Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread |
| title_full_unstemmed | Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread |
| title_short | Forecasting infections with spatio-temporal graph neural networks: a case study of the Dutch SARS-CoV-2 spread |
| title_sort | forecasting infections with spatio temporal graph neural networks a case study of the dutch sars cov 2 spread |
| topic | epidemic prediction deep learning spatio-temporal graph neural networks real world evidence COVID-19 |
| url | https://www.frontiersin.org/articles/10.3389/fphy.2023.1277052/full |
| work_keys_str_mv | AT vmaximecroft forecastinginfectionswithspatiotemporalgraphneuralnetworksacasestudyofthedutchsarscov2spread AT sennacjlvaniersel forecastinginfectionswithspatiotemporalgraphneuralnetworksacasestudyofthedutchsarscov2spread AT cosimodellasantina forecastinginfectionswithspatiotemporalgraphneuralnetworksacasestudyofthedutchsarscov2spread AT cosimodellasantina forecastinginfectionswithspatiotemporalgraphneuralnetworksacasestudyofthedutchsarscov2spread |