Huber, MeikeMeikeHuberAgarwal, DhruvDhruvAgarwalSchmitt, RobertRobertSchmitt2023-09-072023-09-072023https://publica.fraunhofer.de/handle/publica/45038510.1108/IJQRM-09-2022-02682-s2.0-85148943031Purpose: The determination of the measurement uncertainty is relevant for all measurement processes. In production engineering, the measurement uncertainty needs to be known to avoid erroneous decisions. However, its determination is associated to high effort due to the expertise and expenditure that is needed for modelling measurement processes. Once a measurement model is developed, it cannot necessarily be used for any other measurement process. In order to make an existing model useable for other measurement processes and thus to reduce the effort for the determination of the measurement uncertainty, a procedure for the migration of measurement models has to be developed. Design/methodology/approach: This paper presents an approach to migrate measurement models from an old process to a new "similar" process. In this approach, the authors first define "similarity" of two processes mathematically and then use it to give a first estimate of the measurement uncertainty of the similar measurement process and develop different learning strategies. A trained machine-learning model is then migrated to a similar measurement process without having to perform an equal size of experiments. Similarity assessment and model migration Findings: The authors’ findings show that the proposed similarity assessment and model migration strategy can be used for reducing the effort for measurement uncertainty determination. They show that their method can be applied to a real pair of similar measurement processes, i.e. two computed tomography scans. It can be shown that, when applying the proposed method, a valid estimation of uncertainty and valid model even when using less data, i.e. less effort, can be built. Originality/value: The proposed strategy can be applied to any two measurement processes showing a particular "similarity" and thus reduces the effort in estimating measurement uncertainties and finding valid measurement models.enMachine learningMeasurement uncertaintyModel migrationSimilarity assessmentjournal article