Integrating machine learning and system dynamics

Tools

Mahmoud, Hesham and Korzilius, Hubert and Olde Rikkert, Marcel and Howick, Susan and Rouwette, Etienne and Schoenberg, William; Hoorani, B and Rouwette, Etienne, eds. (2025) Integrating machine learning and system dynamics. In: Methodology to Address Grand Challenges: Theory, Practice, and Interventions. Edward Elgar Publishing. (In Press)

[thumbnail of Mahmoud-etal-EE-2025-Integrating-machine-learning-and-system-dynamics] Text. Filename: Mahmoud-etal-EE-2025-Integrating-machine-learning-and-system-dynamics.pdf
Accepted Author Manuscript
Restricted to Repository staff only until 1 January 2099.
Download (1MB) | Request a copy

Abstract

The exponential data growth across industry and academia has made machine learning (ML) a promising tool to support decision-making in public and private organizations. The rise of ML to explore correlations in multidimensional data provides a remarkable opportunity for handling big data. However, ML approaches still receive skepticism for both the opacity and complexity of the analysis process which represent obstacles to the wider use of ML. However, system dynamics (SD) modeling, and in particular its participatory version, could complement ML. By involving stakeholders and domain experts in developing conceptual or computational models, data construction can be sped up and the analysis process can be made more transparent. One of the grand challenges worldwide that could benefit from such a sophisticated methodological approach is the cost-effective management of a limited workforce in the elderly healthcare system to support a rapidly growing older population with care needs. The Netherlands, with its aging population, also faces unprecedented pressure on its health- and social care systems as over 70% of the senior citizens with diseases such as dementia live at home and are cared for by their family or the surrounding community, while there increasingly is an insufficient number of professionals available to support these caregivers or take over their care at end stages of dementia. We use this problem as a showcase to demonstrate the combination of participatory SD and ML, and discuss strengths, limitations, and suggestions for implementation in practice and science.

ORCID iDs

Mahmoud, Hesham, Korzilius, Hubert, Olde Rikkert, Marcel, Howick, Susan ORCID logoORCID: https://orcid.org/0000-0002-0796-7981, Rouwette, Etienne and Schoenberg, William; Hoorani, B and Rouwette, Etienne
CORE (COnnecting REpositories)