guest ::
%0 Thesis %A Jungnickel, Robert %T Integrating human knowledge into a machine-learning process for modelling ongoing transitions in socio-technical systems %I Rheinisch-Westfälische Technische Hochschule Aachen %V Dissertation %C Aachen %M RWTH-2025-03106 %P 1 Online-Ressource : Illustrationen %D 2025 %Z Veröffentlicht auf dem Publikationsserver der RWTH Aachen University %Z Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2025 %X Complex systems, such as fundamental transition in socio-technical systems, demand that decision-makers move beyond simple historical analysis to achieve a comprehensive understanding. The vast amount of information – whether generated by humans or through machine-learning – can lead to cognitive and analytical limitations. To address these challenges, the objective of this work is to develop a transition model as a network that integrates human knowledge into a machine-learning process to model and predict the relationships between actors in ongoing transitions. The main outcome of this work is that the integration of qualitative human knowledge into a quantitative machine-learning process can provide measurable insights into ongoing socio-technical transitions. This is demonstrated through the development of a transition model that comprises two interrelated sub-models: the visual network model and the human-centric machine-learning (HCML) model. The work started with theoretical considerations on how to methodologically integrate quantitative and qualitative approaches to enable the monitoring of ongoing transitions. While the visual network model provided the foundational dataset of transition actors and the links between them, the HCML model facilitated ongoing human-machine interaction to explore the prediction of links between these actors. By simulating human feedback, the model optimised its predictions and demonstrated its potential to predict behaviour during ongoing transitions. The ongoing transition from fossil fuels to alternative fuels in the European market served as a test laboratory to identify practical requirements and strategies for the transition model’s development. The Cluster of Excellence ‘The Fuel Science Centre (FSC)’ at RWTH Aachen University represented the use case to evaluate the model development. In conclusion, this work extends existing approaches to transition modelling and temporal link prediction by incorporating human knowledge into an ongoing machine-learning process. The developed transition model is general applicable and serves as a foundational reference for capturing the complex behaviour of socio-technical transitions, whether in Europe or beyond. %F PUB:(DE-HGF)11 %9 Dissertation / PhD Thesis %R 10.18154/RWTH-2025-03106 %U https://publications.rwth-aachen.de/record/1008710