Automated model discovery based on COVID-19 epidemiological data from Thuringia, Germany
6 affiliations
- 1School of Computing and Communications, Lancaster University Leipzig, Leipzig, Saxony, Germany
- 2Leibniz Institute of Photonic Technology, Jena, Thuringia, Germany
- 3Institute of Inorganic and Analytical Chemistry, Friedrich Schiller University Jena, Jena, Thuringia, Germany
- 4Department of Anaesthesiology and Intensive Care, Jena University Hospital, Jena, Thuringia, Germany
- 5Leibniz Centre for Photonics in Infection Research (LPI), Jena, Thuringia, Germany
- 6Institute of Physical Chemistry and Abbe Center of Photonics, Friedrich Schiller University Jena, Jena, Thuringia, Germany
Abstract
The COVID-19 pandemic presented severe challenges in understanding and predicting the spread of infectious diseases, necessitating innovative approaches beyond traditional epidemiological models. This study introduces an advanced method for automated model discovery using the Sparse Identification of Nonlinear Dynamics (SINDy) algorithm, leveraging a dataset from the COVID-19 outbreak in Thuringia, Germany, encompassing more than 400,000 patient records and vaccination data. We develop a flexible, data-driven model capturing pandemic dynamics, incorporating external factors and interventions into the mathematical framework. The fixed coefficient values globally determined by SINDy were not accurate for local modelling of the data, a limitation of prior SINDy-based epidemiological applications that our framework directly addresses. We therefore refined our technique based on the differential equations as found by SINDy, by investigating three modifications that account for recent local data, each offering different strengths for short-term prediction, scenario analysis, and capturing nonlinear effects, achieving an R² of 0.87 at a one-week horizon. In a first approach, we re-optimized the coefficient values using seven days of past data, without changing the globally determined differential equation. In a second approach, we allowed a temporal dependence of the coefficient values, fitted using all previous data, in combination with regularization. As a last method, we kept the coefficients fixed to the original values but augmented the differential equation with a small neural network, locally optimized to the data of the past week. Our results link vaccination and public health measures to the pandemic's trajectory. The proposed model allows simulation of intervention scenarios, such as vaccination strategies and public health interventions. While the current study is based on retrospective data from a single region, the framework could serve as a basis for exploring responses to future outbreaks, subject to further validation.
Related
- Synthesis-aware generative design in trillion-scale chemical spaces for automated drug discovery
- Scaling genetic discovery through automated post-GWAS interpretation
- Causal discovery and inference in COVID-19: A multiscale longitudinal analysis using mixtures of DAGs
- Epidemiological characteristics and transmission dynamics of dengue in Ningbo City, China before, during, and after the COVID-19 pandemic interventions
- A new semi-automated, motility-based screening assay for discovery of compounds with activity against the juvenile stage of Fasciola hepatica
BibTeX
@article{shareh2026automated,
title = {{Automated model discovery based on COVID-19 epidemiological data from Thuringia, Germany}},
author = {Morteza Babazadeh Shareh and Florian Kleiner and Michael Böhme and Corinna Hägele and Petra Dickmann and Rainer Heintzmann},
journal = {PLOS One},
year = {2026},
month = oct,
doi = {10.1371/journal.pone.0344409},
url = {https://doi.org/10.1371/journal.pone.0344409},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Automated model discovery based on COVID-19 epidemiological data from Thuringia, Germany
AU - Morteza Babazadeh Shareh
AU - Florian Kleiner
AU - Michael Böhme
AU - Corinna Hägele
AU - Petra Dickmann
AU - Rainer Heintzmann
PY - 2026
DA - 2026/10/01
JO - PLOS One
DO - 10.1371/journal.pone.0344409
UR - https://doi.org/10.1371/journal.pone.0344409
AB - The COVID-19 pandemic presented severe challenges in understanding and predicting the spread of infectious diseases, necessitating innovative approaches beyond traditional epidemiological models. This study introduces an advanced method for automated model discovery using the Sparse Identification of Nonlinear Dynamics (SINDy) algorithm, leveraging a dataset from the COVID-19 outbreak in Thuringia, Germany, encompassing more than 400,000 patient records and vaccination data. We develop a flexible, data-driven model capturing pandemic dynamics, incorporating external factors and interventions into the mathematical framework. The fixed coefficient values globally determined by SINDy were not accurate for local modelling of the data, a limitation of prior SINDy-based epidemiological applications that our framework directly addresses. We therefore refined our technique based on the differential equations as found by SINDy, by investigating three modifications that account for recent local data, each offering different strengths for short-term prediction, scenario analysis, and capturing nonlinear effects, achieving an R² of 0.87 at a one-week horizon. In a first approach, we re-optimized the coefficient values using seven days of past data, without changing the globally determined differential equation. In a second approach, we allowed a temporal dependence of the coefficient values, fitted using all previous data, in combination with regularization. As a last method, we kept the coefficients fixed to the original values but augmented the differential equation with a small neural network, locally optimized to the data of the past week. Our results link vaccination and public health measures to the pandemic's trajectory. The proposed model allows simulation of intervention scenarios, such as vaccination strategies and public health interventions. While the current study is based on retrospective data from a single region, the framework could serve as a basis for exploring responses to future outbreaks, subject to further validation.
SN - 1932-6203
ER -