In the Stroke DynamiX project, scientists are developing data-driven methods for modelling and predicting stroke progression using machine learning techniques.
Two scientific publications from the project have recently been published: Computers in Biology and Medicine presents an approach to mixed-effects additive Bayesian networks that structurally analyses multicentre clinical data and explicitly takes cross-centre heterogeneity into account. Another article in Mathematics provides the theoretical foundations for additive Bayesian networks in a mathematical context.
These publications show how statistical learning methods enable dynamic, patient-oriented models in stroke research.
Delucchi, A., et al. (2026). Mixed-effects additive Bayesian networks for the assessment of ruptured intracranial aneurysms: Insights from multicenter data. Computers in Biology and Medicine, 201, 111380. DOI: 10.1016/j.compbiomed.2025.111380
Champion, T., Delucchi, A., & Furrer, R. (2025). All for one or one for all? A comparative study of grouped data in mixed-effects additive Bayesian networks. Mathematics, 13(22), 3649. DOI: 10.3390/math13223649
The abn package used for the analysis is available on CRAN and GitHub.