04/2021 – 10/2021
Large sets of process, meta and quality data are generated in the surface mount technology (SMT) of electronics manufacturing. Especially in research, this data is also used to train machine learning (ML) based models for predictive maintenance, process monitoring, quality control and prediction. Re-training these models or re-developing models is time consuming due to the large amount of data involved.
The aim of this work is to develop an ML model for an exemplary selected use case. Subsequently, the influence of the training data set on the model performance with respect to the selected application shall be investigated. Based on this, conclusions will be drawn about the requirements for retraining a model in an industrial environment.